ULTRAVIOLET RADIATION AND THE EYE:

AN EPIDEMIOLOGIC STUDY*

BY Hugh R. Taylor, MD

INTRODUCTION
ALL OF US ARE EXPOSED TO SOME DEGREE OF SUNLIGHT. THE AMOUNT OF
exposure can vary greatly among different occupations and different
recreational activities. The study described here was undertaken to see if
sun exposure was harmful to the eye. Specifically, it examined possible
associations between levels of exposure to ultraviolet radiation (UVR) and
the formation of cataract, macular degeneration, and corneal disease.
As a visual organ, the eye is very much affected by light. All day,
thousands of molecules of rhodopsin are altered by visible light, but with
ordinary exposure these light-induced changes are short-lived and rapidly
reversed. However, intense exposure to either the broad band of visible
light or to narrower specific bands in the visible spectrum, such as those
produced by a laser, can cause permanent ocular damage. For example,
the occurrence of retinal burns in eclipse blindness is well known,' and
retinal laser photocoagulation is one of the major advances in the treat-
ment of eye disease of the last two decades.2
Not all bands of eletromagnetic radiation emanating from the sun are in
the visible spectrum, and many of the nonvisible bands can have a serious
impact on biological function. While most harmful solar radiation is
filtered out by the atmosphere, the sunlight that does reach the earth's
surface contains sufficient amounts of UVR to cause sunburn3 and a
variety of skin cancers.4
For many years, it has been suggested that exposure to sunlight (or,
more specifically, UVR) may be associated with an increased risk of senile
cataract5 and possibly even with senile macular degeneration (SMD; now
also referred to as age-related or aging-related macular degeneration or
maculopathy). 1 Most of the initial suggestions concerning the association
between UVR and cataract came from astute observations by experienced
physicians5 rather than rigorous epidemiologic studies. More recent field
*From the Dana Center for Preventive Ophthalmology, the Wilmer Ophthalmological
Institute, the Johns Hopkins University Schools of Medicine and Public Health, Baltimore,
Maryland. Supported in part by NEI grant EY-06547 and in part by NIH grant EY04547.
TR. AM. OPHTH. Soc. vol. LXXXVII, 1989
 UV and the Eye 803

studies, reviewed in greater detail below, also suggest an association.

However, all of these studies suffer from a lack of precision both in the
definition of cataract and in the determination of UVR exposure. They
have used a variety of definitions of cataract and often have failed to
distinguish between different types of cataract. None has attempted to
determine the exposure of each individual to sunlight or UVR. Instead,
they have taken the amount of radiation in the region in which the person
lived as the indicator of personal lifetime exposure, assuming that every-
one in a given region will have a similar exposure to UVR. This results in
an epidemiologic trap known as the "ecology fallacy," since different
people in the same environment differ in their individual exposures. As a
result, the suspected relationship between UVR exposure and cataracts
has awaited rigorous scientific evaluation. The case for an association
between UVR exposure and SMD is even less clear. The following study
was undertaken in an attempt to answer these questions definitively.
The following introductory sections review the basis for the notion that
UVR may cause ocular damage. They include a review of the photobiol-
ogy of UVR. The biochemical changes associated with cataract are dis-
cussed with emphasis on the potential role for UVR. Finally, the epidemi-
ologic studies of UVR and cataract are evaluated. This body of evidence
suggests, overall, a physical and biochemical basis for an association
between UVR and cataract, although to date, the epidemiologic evidence
has been unconvincing. The review will also examine the possibility of an
association between UVR and SMD.

DESCRIPTION OF ULTRAVIOLET RADIATION

Physical Definition of UVR

The spectrum of nonionizing radiation ranges from short wavelength
UVR (wavelength 100 nm) through to far infrared radiation (1 mm or
100,000 nm).6 Ionizing radiation has wavelengths shorter than 100 nm and
includes x-rays, gamma rays, and cosmic rays. Hertzian radiation has
wavelengths greater than 1 mm and includes radar, microwave radiation,
and radio waves, which can be up to several kilometers in length. The
visible spectrum lies between 400 nm (indigo) and 760 nm (red). Above
the visible spectrum is infrared radiation, and below the visible spectrum
are the shorter wavelengths of nonionizing radiation called UVR. Much of
the nonionizing radiation is absorbed by the earth's atmosphere and does
not reach the earth's surface.6 Wavelengths below 290 nm are totally
absorbed by the ozone layer in the stratosphere, and longer wavelengths
are absorbed to a lesser extent. Thus, in nature, one does not encounter
804 Taylor

UVR below 290 nm, although the physical spectrum of UVR ranges from
100 nm to 400 nm.
UVR has been subdivided into three bands: UV-A (400 to 320 nm), UV-
B (320 to 290 nm), and UV-C (290 to 100 nm). This arbitrary subdivision is
based on the biologic effects ofthe different wavelengths or bands.7 UV-A,
or near UV, produces sun tanning (the browning of the skin due to an
increase in the skin content of melanin). UV-A is also responsible for
photosensitivity reactions. UV-A is commonly encountered and is emitted
by so-called black lights, which are often used to make objects fluoresce
and are also used in tanning salons. UV-B is the sunburn spectrum and
causes sunburn (painful erythema) and tissue damage (blistering). UV-B is
associated with skin cancer. 4-8 UV-C is germicidal and may also cause skin
cancer. UV-C, or far UV, is not commonly encountered on the earth's
surface and comes entirely from artificial sources such as germicidal UV
lamps or arc welding. Although UVR is only 5% of the sun's energy, it is
the most hazardous portion encountered by man. Furthermore, although
UV-B is only 3% of the UVR that reaches the earth's surface, it is much
more biologically active than UV-A.9
Ocular Exposure to UVR
The amount of UVR that reaches the eye can vary enormously. UVR is
scattered across the whole sky by the Rayleigh effect, just as blue light is
scattered.8 Light or broken clouds do not significantly reduce the level of
UVR, although levels are reduced by heavy cloud cover.8 A sky with a
clear horizon for 3600 provides for a maximal exposure; when hills, trees,
or buildings obstruct part or all of the horizon, the UVR exposure is
reduced proportionally.'0 UVR can also be reflected by the ground, the
amount depending greatly on the type of surface. Grass and soil reflect
only 1% to 5% of UV-B, water 3% to 13%, sand and concrete about 7% to
18%, and fresh snow up to 88%. 10
The eye is protected and shielded from UVR by a number of factors""12
and only receives a small fraction of ambient UV-B under normal circum-
stances. The normal horizontal alignment of the eye and the orbit signifi-
cantly reduces ocular exposure to whole-sky irradiation. Further anatomic
protection is provided by the brows, the nose, and the cheek. 13 The eye is
relatively unprotected laterally, although the transmission of UVR by
internal reflection in the cornea may lead to a concentration of UV
irradiation at the nasal limbus.'4 The eyelids provide protection that is
further enhanced by squinting, a common reflex in bright sunlight.'3
Other factors that can influence ocular UVR exposure in a given environ-
ment include wearing a hat and the use of eyeglasses. 1112 Taken together,
 UV and the Eye 805
these different factors result in an ocular exposure that is considerably less
than the ambient UVR level. The most important factors in determining
an individual's exposure in a given environment are whether protective
glasses or a hat are worn.12
Ocular Transmission of UVR
Not all the UVR that reaches the eye passes back to the retina-a fact with
important photobiological implications. The amount of radiation that is
absorbed determines the potential for damage to the absorbing tissue.
Energy from absorbed radiation must be dissipated, and it is this dissipa-
tion that results in damage. Radiation that is not absorbed by a superficial
tissue will be transmitted and can affect a deeper tissue.
The cornea absorbs almost 100% of UV-C radiation (below 290 nm), but
transmission rapidly increases for longer wavelengths, so that, for in-
stance, 60% of radiation at 320 nm is transmitted by the cornea. 1&18 The
normal, young human lens absorbs most UVR below 370 nm. With age,
the human lens yellows and absorbs even more UV-A and also absorbs
shorter visible wavelengths.18'19 In adults, less than 1% of radiation
between 320 nm and 340 nm and only 2% of radiation of 360 nm reaches
the retina.20 The pattern of absorption shown in the classic series of
transmittance curves published by Boettner and Wolter17 indicates that
the lens is exposed to and absorbs most of the UV-B that reaches the eye.
Mechanisms of Phototoxicity
The mechanisms of phototoxicity are complex and not totally understood.
All electromagnetic radiation exhibits both wave-like (oscillatory) and
particle-like (photon) characteristics.6 The energy carried by a photon is
directly proportional to its frequency, thus the shorter the wavelength,
the higher the energy. The energy of a photon is absorbed by the atom or
molecule with which it collides. Low-energy infrared photons will carry
enough energy to affect the rotational or vibrational state of an atom or a
molecule and can produce warming.7 The higher energy UVR photons,
however, can alter the energy state of the electrons, making the atom or
molecule electronically excited and, therefore, relatively unstable. This
instability can lead to chemical reactions including photo-oxidation. Even
higher energy photons such as gamma rays can cause an electron to be
removed entirely from the molecule thereby causing ionization.
The radiant energy of UVR can be absorbed by nucleic acids, proteins,
or other molecules within the cell. Some energy may be dissipated as
heat, but an excited molecule may be structurally altered or cleaved or it
may react with other molecules by forming new bonds. The capacity of a
given atom or molecule to absorb radiant energy is dependent on its
806 Taylor

physicochemical properties, and the characteristics of a tissue are in turn

dependent on the properties of its constituents. The lens proteins are rich
in the amino acid residues of tryptophan, tyrosine, and phenylalanine;
and these proteins absorb most of the radiant energy below 300 nm.
Other chromophores and pigments in the lens appear to absorb most of
the energy in the 300 nm to 400 nm range.2'

Ocular Action Spectrum for UVR

The eye is much more sensitive to damage by some wavelengths than
others. The term "action spectrum" can be defined as the amount of
irradiation at a given wavelength or band of wavelengths that is sufficient
to cause damage. It is the threshold level for damage and is specific for
different tissues. In their classic work, Verhoeff et al22 showed that
repeated subliminal exposures delivered within minutes or hours to the
eye are additive for up to 24 hours. For suprathreshold corneal damage, at
least, there is a period of latency between exposure and evidence of
damage which varies with dose but can range from 30 minutes to 24
hours. The latency period accounts for the delay in presentation of weld-
ers with flashburns. For the rabbit lens, a latency period of 5 to 10 days
has been reported for reversible lesions and 2 to 14 months for irrevers-
ible lesions. 16 No data exist for the action spectrum of the human lens or
retina. The action spectrum for the human cornea is generally similar to
that of subhuman primates, although the human cornea is more sensitive
to UVR than is the cornea of either monkeys or rabbits.23 The most active
waveband for corneal damage in both monkeys and rabbits is 270 nm.
Exposure of rabbit or guinea pig lens to wavebands between 297 nm and
365 nm results in cataract, with a peak sensitivity shown at 297 nm; i'n
monkeys, the sensitivity of the lens falls rapidly for radiation above 313
nm.16 24The rabbit lens is 100 times more sensitive to 320 nm radiation
than to 365 nm radiation.Y5 Thus, these studies have shown that the lens is
particularly susceptible to damage from UV-B.
Ham and co-workers26 have studied the action spectrum for retinal
injury in aphakic primate eyes. They have found that at 325 nm, the
shortest waveband they tested, the retina was six times more sensitive
than at 441 nm. UVR produced early and irreparable damage to the rods
and cones with later changes in the retinal pigment epithelium. The later
changes were similar to those seen with blue light (441 nm), although blue
light did not cause photoreceptor damage.27 These authors emphasized
the extreme sensitivity of the primate retina to UVR-induced damage.
 UV and the Eye 807
BIOCHEMICAL CHANGES IN CATARACT
This section reviews what is known of the biochemistry of cataract and the
next section shows how these changes could be related to UVR exposure.
Changes in Lens Proteins
The lens has an extraordinarily high protein content, and changes in the
lens proteins result in cataract. Overall, proteins form 35% of the wet
weight of the lens, and in the nucleus, they form 65% to 70%.28 In the
young lens, much of this protein is in a soluble form; but with aging, there
is an increase in the amount of water-insoluble protein.29 With cataract,
there is an even more marked increase in the proportion of water-insol-
uble protein and a decrease in the amount of water-soluble protein.30
Benedek3l suggested that the opacification seen in cataract was due to
the formation of large protein aggregates. These aggregates are formed by
the cross-linkage of soluble proteins to form larger, irregular, and insolu-
ble protein masses which then scatter light and cause the lens to appear
opaque. These protein aggregates are stabilized in part by covalent disul-
fide bonds32 which result from the oxidation of exposed thiol groups.33.34
Some of these aggregates are associated directly with the membranes of
the lens fiber.28,35
There are differences in the types of protein aggregates found in
nuclear and cortical cataracts.28,36 Nuclear cataract is characterized by
high levels of insoluble protein complexes that are covalently linked to the
lens fiber membranes. These protein complexes may be derived from
gamma crystallin.37 In the cortex, however, high molecular-weight aggre-
gates of insoluble protein are not found. Spector28 has suggested that in
cortical cataract there may be a complete rupture of the protein-mem-
brane complex and that the protein lumps then disaggregate from the
ruptured membrane remnants.
Most of the changes in lens proteins can be explained either by oxida-
tive28 or by osmotic changes.38 With normal aging, there is a slow oxida-
tion of the thiol groups of the cysteine and methionine residues in the lens
fiber membranes. In cataract there is a more marked oxidation in the
membrane, and the crystallins inside the fiber are also oxidized.39 Oxida-
tion of a single thiol group in gamma crystallin causes an unfolding of the
molecule, which leads to the exposure of more groups for oxidation and a
compounding of the damage.28 The oxidized thiol groups can form cova-
lent bonds, which lead to the aggregation of lens proteins to each other or
to the cell membranes. Osmotic swelling leads to progressive disruption
of the lens fiber, and this has been extensively studied in the context of
"sugar" cataract.38,40 Initially, there is an increase in the intracellular
808 Taylor
water with a loss of biochemical constituents. This leads to a marked
swelling of the lens fiber and then to its disintegration with vacuole
formation and opacification.
Changes in the lens proteins can lead to light scattering and opacifica-
tion by a number of different mechanisms.4' Aggregation and the forma-
tion of large water-insoluble proteins can cause light scatter. An increase
in turbidity or syneresis also leads to increased light scatter, and this
mechanism may be important in osmotic cataracts. Other mechanisms
include membrane degeneration, the disorientation of cytoskeletal ele-
ments, and phase separation. The latter is seen in experimental cold
cataract. Because there is essentially no turnover of the lens proteins, any
abnormal or denatured protein will remain, and these altered molecules
will gradually accumulate.
Recently, it has been discovered that lens proteins previously thought
to be inert and to have only a structural role may, in fact, have important
functional significance. Reptilian epsilon crystallin has been found to have
lactate dehydrogenase activity and thus may be an important glycolytic
enzyme.42 Similarities exist between delta crystallin and arginosuccinate
lyase, gamma crystallin and enolase, and Slll crystallin and glutathione
s-transferase.43 It is unclear whether these functional roles are important
in lens physiology or what effect UVR damage might have on this activity.
Changes in Lens Pigmentation
Diurnal animals, including man, have a number of pigments in their
lenses.44 Teleologically, these are thought to provide some protection for
the retina from the potentially harmful effects of UVR and blue light.45
Included among these pigments is a series oflow molecular-weight, water-
soluble compounds related to kynurenine, and oxidative product of tryp-
tophan.46,47 Another is an anthranilic acid derivative that is bonded to the
lens proteins and is also possibly an oxidation product of tryptophan.36
Both of these pigments increase with age and with the proportion of
insoluble proteins.2' Another nontryptophan water-soluble pigment,
which appears to decrease with age, has also been described.2'
Although much attention has been given to the pigments of the lens in
the biochemical literature of cataract, it is not clear what is the biological
or visual significance of these pigments. The lenses of older people are
browner than those of younger people. This is termed "nuclear sclerosis"
in the strict sense and can be regarded almost as a normal change seen
with aging.48 This increase of nuclear pigmentation (nuclear sclerosis),
however, is independent of nuclear opacification (nuclear cataract).49 Ad-
vanced browning of the lens acts as a filter and reduces the amount of
 UV and the Eye 809

visible light reaching the retina, especially at the blue end of the spec-
trum. However, the major visual disability associated with cataract is due
to light scatter and opacification that is attributable to changes in lens
proteins and not to an increase in lens pigmentation.

ASSOCIATION BETWEEN UVR AND CATARACT

This section examines how UVR could contribute to the biochemical
changes seen in cataract and then reviews the experimental and epidemi-
ological evidence for an association.
Biochemical Basis for an Association Between UVR and Cataract
There are good biochemical grounds to believe that UVR may be an
important cause of at least some types of cataract. UVR could directly
affect the lens and cause cataract by at least four different mechanisms: (1)
photo-oxidation of free or protein-bound tryptophan; (2) photosynthetic
processes involving activated species of oxygen; (3) disruption of the
membrane-cation transport system; and (4) damage to nucleic acids in
lens epithelial cells.
Photo-oxidation of Tryptophan. Lens proteins contain photosensitive
amino acid residues, most notably those of tryptophan. It has been
suggested that UVR-induced photo-oxidation of tryptophan may be a
critical first step that initiates changes in lens proteins.21 In a series of
pioneering experiments, Pirie465'0 and Dilley5l showed that lens proteins
exposed to sunlight turned brown, and these changes were similar to
those found in human lenses. They demonstrated the photo-oxidation of
tryptophan to give N-formyl kynurenine, a brown compound. Subse-
quently, other species of pigments derived from kynurenine were found
by van Heyningen,52 who also showed that these pigments caused more
rapid oxidation of lens proteins.47'53 It has been shown that free tryp-
tophan in the lens can be oxidized54 as can at least a portion of the
protein-bound tryptophan.55 The effects of these molecules and other
photosensitizing compounds on the impact of UVR have been compre-
hensively reviewed. 2128
However, as Harding and Dilley5' point out, the proteins of the brown
cataractous nucleus do not appear to have less tryptophan than normal
lens proteins, and the brown cataract change affects only the nucleus and
not the anterior cortex, which receives the highest levels of UVR.
Activated Species of Oxygen. UVR can generate a series of highly
reactive toxic oxidants by its effect on molecular oxygen. 57 These oxidants
include singlet oxygen, superoxide, and hydrogen peroxide. Specific
810 Taylor

enzymes that destroy these destructive oxygen species include superox-

ide dismutase, which inactivates superoxide, and catalase, which inacti-
vates hydrogen peroxide. These enzymes are found in the lens epithelium
and occur in high levels in the retina. However, they are present only in
very low levels within the lens itself, mainly in the cortex.5 The activities
of both catalase and superoxide dismutase decrease with age, and catalase
is inactivated by exposure to UVR.21,59
Other antioxidants are also found within the lens. The most important
is glutathione, although ascorbic acid (vitamin C) and alpha tocopherol
(vitamin E) are also present.28'58 Reduced glutathione becomes oxidized
as it repairs oxidative damage. Oxidized glutathione is returned to the
reduced state by glutathione reductase. Reduced glutathione is found in
high levels in the lens epithelium and the cortex but in much lower levels
in the nucleus.57,60 The concentration of glutathione in the lens decreases
with age and with exposure to UVR.61 Glutathione may have multiple
roles in maintaining lens clarity by maintaining lens protein and mem-
brane thiol groups in the reduced state as well as scavenging free radicals
and hydrogen peroxide.60
Ascorbate can also repair photo-oxidative damage.57'62 Ascorbate is
normally abundant in the lens, although its levels decrease with age and
in the presence of cataract. Ascorbate protects Na/K ATPase against
UVR63 and inhibits the browning of lens proteins.64 Recently, Bron and
Brown62 suggested that certain cortical opacities, "retrodots," which are
probably oxalate deposits, may result from the oxidation of ascorbate in
the lens.
Oxidants may be formed not only in the lens but also in the aqueous
humor. They could diffuse from the aqueous into the lens and cause
oxidative damage. High levels of hydrogen peroxide have been reported
in the aqueous humor of patients with cataract, and the exposure of the
lens in vitro to hydrogen peroxide leads to cortical opacification.65 Hydro-
gen peroxide affects membrane transport systems, including Na/K ATPase
activity. 66
Disruption of Membrane Cation Transport. Normal Na/K ATPase activ-
ity, the so-called sodium pump, is crucial for lens clarity as it maintains
the osmotic balance within cells. A loss of activity leads to osmotic
swelling of lens fibers, as can be seen when the enzyme is inhibited by
ouabain.38 This mechanism may be important in at least some experimen-
tal congenital cataracts. 67 Na/K ATPase is sensitive to damage by UVR but
can be protected by ascorbate in the lens.63'68 This enzyme is found in
larger quantities in the retina, where it is also sensitive to UVR inactiva-
tion.21
 UV and the Eye 811

Damage to Lens Epithelium. Finally, UVR could damage the lens

epithelium by a direct effect on deoxyribonucleic acid (DNA). UVR can
damage DNA in the skin.68 UVR exposure of the lens of animals leads to
histologic changes in the lens epithelium that are similar to those seen in
skin. 69,70 The lens epithelium shows a lack of differentiation which may be
due to DNA damage. The undifferentiated cells migrate to the posterior
lens surface and cause posterior opacities. These changes are similar to
those seen with ionizing radiation in man and human posterior subcapsu-
lar opacities.7' Apart from a direct effect on the differentiation of lens
epithelium, UVR damage to the lens epithelium could have indirect
effects on the lens by altering lens nutrition or by altering the metabolic
and protective role of the epithelium.
Experimental Basis for an Association Between UVR and Cataract
Experiments in animals suggest that UVR may cause cataract. Cortical
and posterior subcapsular cataracts have been induced by UV irradiation
in a number of different experimental animals. Rohrschneider72 exposed
guinea pigs to a mercury vapor lamp (293 nm to 303 nm) and produced
clouding of the anterior cortex. Bachem'6 confirmed these findings in
both guinea pigs and rabbits, using a number of different light sources,
and determined the action spectra. Zigman and co-workers73 reported
subcapsular and punctate cortical opacities in albino mice that had been
exposed to broadband UVR from 40 W blacklight lamps (300 to 400 nm).
The mice were exposed for 12 hours a day. After 35 weeks of exposure,
changes were seen on slit lamp examination, and after 60 weeks of
exposure, clear-cut cortical opacities were seen on histologic examination.
At this time, the posterior migration of undifferentiated lens epithelium
was also noted.69 Pitts and co-workers24 produced cortical and posterior
subcapsular opacities in pigmented rabbits exposed to short-duration
exposures to UVR between 295 nm and 315 nm. They commented
specifically on the low radiant exposures of UV-B needed to produce these
lenticular opacities. Keeney and Rapton74 exposed nonpigmented mice to
daylight and caused extensive ear damage and corneal changes, but lens
opacities were not seen.
Permanent lens opacities can be induced in vivo in laboratory animals
with either acute high-dose (suprathreshold) or chronic low-dose (sub-
threshold) exposure to UVR, specifically UV-B. The opacities are either
cortical or subcapsular. Nuclear opacities and brunescent changes have
not been induced in these experiments. The lack of brunescent change
could be because these animals do not have pigment (chromophores) in
their lenses.
812 Taylor

Epidemiologic Basis for an Association Between UVR and Cataract

Epidemiologic and clinical observations also suggest a link between sun-
light exposure and cataract. Cataracts occur more commonly in tropical
areas than in more temperate regions. 575 This is often ascribed to greater
sun exposure, but few epidemiologic studies have examined, even indi-
rectly, the role of sunlight in cataractogenesis. There is a greater preva-
lence of cataracts in the sunnier part of Romania,76 and cataracts are more
common in Israel than in England.77 People having cataract surgery are
more likely to have brunescent cataracts if they live closer to the equator
or work outdoors.78 People in the United States who live more than half
their lives in areas with high sunlight or UVR levels have a greater risk of
cataract. 79,80 This association holds for cortical cataract but not for nuclear
cataract. 81
An earlier study in Australian aborigines showed an epidemiologic as-
sociation between the occurrence of senile cataract and resident sunlight
exposure and, in particular, local levels of UV-B radiation.82 The associa-
tion was specific, consistent, and showed a dose-response relationship.
Hollows and Moran83 also confirmed and extended these findings. They
found a strong positive correlation between the intensity of UV-B radia-
tion in the zone of residence and the presence of cataract when they
analyzed data from 64,307 aborigines.83 These two studies on Australian
aborigines are important for several reasons. First, aborigines receive
very little protection from the environment because they spend virtually
all of their time living in the open, so their individual exposure is likely to
be more closely correlated with the ambient level. Second, the studies
covered a large geographic area with a wide range of environmental
parameters and ambient UVR levels. Third, the study area was lightly
populated, and it was possible to examine almost the entire aboriginal
population living in the study areas rather than a sample. This avoided the
necessity of sample selection with the attendant potential sources of bias.
An association was not found for the Europeans examined by Hollows and
Moran.83 This is probably because of selection bias and lack of quantifica-
tion of individual exposure.
A study of 125,279 Chinese in seven rural areas found that cataract was
more common in areas with more sunlight, and consequently more UV-B
radiation, especially areas at higher altitudes.84 A recent small-scale, case-
control study in Tibet showed greater risk of senile cataract in those who
worked outside for more than 6 hours a day.85 A country-wide survey of
Nepal in which 30,565 lifelong residents were examined also found a
positive correlation between sunlight and cataract.86 In this study, the
investigators had to contend with the marked influence of neighboring
 UV and the Eye 813

mountains on sun exposure, and they came up with a simple instrument

to measure the height of the surrounding mountain mass. An earlier study
in north India had also reported a higher prevalence of cataract in the
plains than in the mountains, which had been taken as proof against an
association between UVR and cataract.87 This earlier study did not at-
tempt to ascertain either personal or ambient UVR exposure and was
confounded by so many other factors that it is hard to interpret.
Although these studies suggest an association between exposure to
sunlight, or UVR, and cataract, the studies share several common short-
comings. None of them have prospectively taken into account the poten-
tial for UVR to preferentially affect one part of the lens. The outcome they
have used was "cataract," often advanced or mature cataract. Sometimes
the cataracts are referred to as being brunescent, but it is unclear whether
they were pure brown nuclear cataracts or whether they were mature
mixed cataracts which happened to be brown.
Furthermore, each of these major studies was "ecological" in that the
exposures of individuals to sunlight or UVR were not ascertained. Rather,
the characteristics of the area of residence, such as latitude or more
refined meteorological data, were used to estimate exposure. This gives a
very imprecise measure of actual individual exposure. It is also possible
that the observed geographic or ecological differences in the rate of
cataract might be due to other confounding factors such as genetic charac-
teristics,88,89 nutritional status,87,90 other climatic factors such as tem-
perature or humidity,82'87 the presence of endemic diseases such as
chronic diarrhea,91'92 or even the availability or utilization of health
services.
In an attempt to control for some of these confounding variables, self-
selected samples of Maryland watermen and their wives and Pennsylva-
nian underground coal miners and their wives were examined.93 Because
of intergroup differences, only limited between-group comparisons could
be made. The watermen could be grouped by their work-related UVR
exposure. Cortical opacities were seen more commonly in those with a
higher UVR exposure history than in those with a lower exposure. No
association was seen with nuclear opacities. SMD as determined by
fundus photography also appeared to be more common in those watermen
under 60 years of age who had a higher UVR exposure, but no trend was
seen in those over the age of 60 years.
In an effort to better quantify personal lifetime UV-B exposure, another
study examined 212 men who had had a skin biopsy taken from the facial
area.94,95 Each biopsy was graded histologically for actinic elastosis, a
marker for cumulative UV-B exposure. A positive, but modest, association
814 Taylor
was found between actinic elastosis and cortical opacities. This association
was strongest in those under age 55. No association between UVR and
nuclear opacities was found.
Despite their differences, these epidemiologic studies suggest an asso-
ciation may exist between UVR and either mature (brunescent) cataract or
cortical cataract in man. None of them suggest an association with nuclear
cataract as such.
ASSOCIATION BETWEEN UVR AND MACULAR DEGENERATION
Much less is known about the role of chronic UVR exposure and macular
damage. Although the lens filters out most of the radiation below 370 nm,
the retina has been shown to be particularly sensitive to any UVR that
does reach it.27,9697 Light can damage the retina by thermoacoustic
damage (as with YAG laser treatment), by thermal damage or burn (as
with xenon photocoagulation), or by photochemical damage. These mech-
anisms have been recently reviewed in detail by Guerry and co-workers. 1
Blue light and UVR usually cause retinal damage by photochemical
means. 1,96 The processes involved are essentially similar to those outlined
for damage to the lens. The retina differs from the lens in that it has a
more active metabolism. It has much higher levels of oxygen and higher
levels of superoxide dismutase and the other antioxidants (especially
ascorbic acid and alpha tocopherol). Furthermore, in the retina, damaged
proteins can be replaced by normal cellular repair mechanisms.
Short, intense exposure to sunlight leads to retinal burns known as
eclipse blindness or solar retinopathy-probably the result of photo-
chemical damage to the photoreceptors.96 Similar damage may be caused
inadvertently by ocular instruments.97,98 Experimentally, the histologic
changes induced by cumulative photic injury in rhesus monkeys resemble
those seen in SMD in man.99
Although much attention has been given to the retinal effects of short,
intense light exposures, the effect of long-term sunlight or UV irradiation
on the retina is unclear. Animal studies by Ham26 and others96 "00' 01 have
demonstrated that longer wavelength UVR and blue light can cause
retinal damage at light levels below those that cause photocoagulation.
Damage from repetitive exposures may be additive. 102,103 However, epi-
demiologic studies have not adequately addressed the question of wheth-
er exposure to high levels of UVR increase the risk of SMD, although the
possibility of such an association has been suggested. 103104 As mentioned
above, a small study of Chesapeake Bay watermen found a weak associa-
tion between UVR exposure and SMD in watermen under the age of 60
years but no association in those over this age.93
 UV and the Eye 815

Some investigators have found SMD to be associated with light iris and
hair color. 105 It has been suggested that this may reflect a lower level of
melanin and thus less protection against photo-oxidative damage because
of the direct absorption of light.105 A small case-control study reported
that the degree of dermal elastotic degeneration in sun-protected skin was
predictive of exudative maculopathy.106 Such information suggests that
individuals whose elastic fibers are more susceptible to photic or other
degenerative stimuli may also have an increased risk of developing SMD.
As discussed above, the aging lens becomes increasingly brown and
acts as a blue-light filter to protect the retina from shorter wavelengths.
Eyes with nuclear sclerosis have been reported to have less SMD than
eyes without nuclear sclerosis.107 This would support the notion of pro-
tection, although it is possible that cases of SMD have been missed in
eyes with nuclear sclerosis. In the same analysis, eyes with cortical
opacities were at greater risk of having SMD, which could suggest a
common cause for these two conditions. Thus, there is some indirect
evidence to suggest a linkage between UVR exposure and SMD.
SUMMARY
The preceding review indicates that an association between UVR and
ocular damage, including cataract, could be suspected on photobiological,
biochemical, experimental, and epidemiologic grounds. There is evi-
dence to suggest that sufficient UVR may reach the lens to cause damage.
In vitro UV exposure can lead to oxidation and denaturation of lens
proteins by a number of different pathways, and the changes induced by
UVR are similar to those seen in human cataractous lenses. There is
experimental evidence in animals that in vivo exposure to UVR can cause
cataracts. Epidemiologic evidence in man also suggests that high levels of
UVR may be associated with cataract. Although intense light exposure is
known to cause retinal damage in animals and man, there is only equivo-
cal epidemiologic evidence to suggest an association between macular
disease and UVR.
The establishment of an adverse effect on UVR on the eye would have
tremendous public health importance. The ocular diseases attributed to
UVR are of major significance in terms of the absolute numbers of people
involved. In the United States alone, over 1 million cataract operations
are performed each year,108 and worldwide, over 17 million people are
blind from cataract. 109 Although SMD affects fewer people than cataract,
it is the leading cause of new cases of blindness in those over age 65 in the
United states. 110 Furthermore, blindness from SMD is generally irrevers-
ible, in sharp contrast to that caused by cataract.
816 Taylor

Protecting the eye from UVR has become a multimillion-dollar-a-year

industry with the development of UVR-absorbing spectacles, sunglasses,
intraocular lenses, and, most recently, contact lenses. Furthermore, there
are strong reasons to believe that we may be facing significantly higher
levels of UVR because of progressive changes in the earth's atmo-
sphere. 111-113 Recent data indicates that chlorofluorocarbon compounds
are causing a significant reduction in the ozone layer in the strato-
sphere.14 The ozone layer is the main atmospheric filter of UVR. It is
therefore of great importance to examine in detail the association between
UVR and ocular damage, specifically senile cataract and SMD.
We have undertaken a definitive epidemiologic study in which we tried
to determine accurately the cumulative UVR exposure of each individual
and correlated this with the presence and severity of the major types of
senile cataract and SMD.
The specific goals of this study were:
1. To identify a study population that had a range of occupational
exposure to UV-B radiation.
2. To determine the annual UV-B exposure for each year of adult life
for each member of the study population.
3. To determine the presence and severity of lens and macular changes
for each member of the study population.
4. To conduct a statistical analysis, assessing the association of UV-B
exposure with senile cataract and with SMD.

METHODS

SAMPLE SELECTION

Baseline Population
The people enrolled in this study were all watermen who work on the
Chesapeake Bay. Every commercial waterman who wants to fish or collect
seafood legally in the state of Maryland must purchase the appropriate
license every season through the Maryland Department of Natural Re-
sources (DNR). For each license applicant, the DNR records the name,
birthdate, address at the time of application, type of license, and date of
application. This information has been coded and filed on data tape for the
last 10 years. The DNR provided us with a copy of these data so we could
define our baseline population.
To be eligible for this study, a subject had to be a male over the age of
30 years who resided in either Somerset County (excluding Princess
Anne) or lower Dorchester County and who had held at least one type of
 UV and the Eye 817

professional waterman license within the last 10 years. The original list
was updated and clarified with assistance from the Maryland Waterman
Association and its county branches, and from telephone listings, local
service clubs, and post office records, together with home visits.
A total of 1203 watermen met the criteria for inclusion in our survey
and formed the baseline population.
Study Population
Of the 1203 watermen we determined were eligible for the survey, 838
(70%) were interviewed and examined during three intensive periods of
data collection. The periods were timed to correspond to the breaks
between the crabbing and oystering seasons that occur in the early spring
and the early fall when most watermen spend 1 to 3 weeks repairing and
refitting their boats. We hoped that they would be more likely to partici-
pate in the study than if we scheduled the interviews and examinations
when they were working on the water.
The first two periods of data collection were conducted in Somerset
County and Dorchester County, in March and September 1985, respec-
tively. Six hundred twenty-one watermen were examined during this time
and are referred to as "primary attenders." Because of concern for possi-
ble bias in this group, a random 10% sample of those who were nonat-
tenders at that time was chosen to determine if important differences
existed between those who had been examined and those who had not.
Exhaustive recruitment efforts were mounted to examine the 68 water-
men randomly selected who are referred to as the "random sample
group." Members of the random sample were sought during the third
period of data collection from December 1985 to January 1986. Additional
efforts were also made at that time to improve the overall response rate by
actively recruiting more nonattenders to the survey. A total of 176 addi-
tional watermen, excluding members of the random sample, were ex-
amined at this time, and they are referred to as the "secondary attenders."
Of the 68 members of the random sample, 17 (25%) were eventually
found not to meet the inclusion criteria in that they were either nonresi-
dents (9), deceased (6), or have given incorrect birthdates at the time of
registration and were actually less than 30 years of age (2). Of the remain-
ing 51, it was possible to contact 46 (90%), and all agreed at least to be
interviewed. Forty-one (80%) also agreed to have the eye and skin exami-
nations.
In summary, the baseline population as determined from the DNR
records consisted of 1203 watermen who met the inclusion criteria. Of
these, 797 watermen who were either primary or secondary attenders and
818 Taylor

41 in the random sample group had eye and skin examinations; thus the
final study sample consisted of 838 men who were both interviewed and
examined. They represent 69.7% of the determined baseline population.
DATA COLLECTION
Recruitment
Before each period of data collection, the survey was widely publicized in
each community. Local watermen and their wives were also hired to assist
with recruitment and field work. Each eligible waterman was contacted
by mail and given a suggested appointment. If a waterman did not keep
this appointment, he received telephone calls and home visits to encour-
age his participation.
Altogether, 360 members of the baseline population (30%) did not
'participate in the study. It was possible to either contact or identify all but
72 of these (6% of baseline population) and ascribe a reason for their
nonparticipation. Of 288 who were either contacted or identified, 117
(10% of baseline population) refused to participate; a further 101 (8%)
were covert refusals as they indicated their willingness to participate but,
in fact, did not do so; 18 (2%) agreed to be interviewed but refused to be
examined; 32 (3%) were known to be living at their given address but
could not be contacted; 15 (1%) were away; 3 were hospitalized; and 2 had
psychiatric disorders.
Interview Data
An interview was administered to each individual by a trained inter-
viewer who gathered information on the following factors: (a) demograph-
ic and background characteristics, including birthdate, education, resi-
dence, characteristics of freckling and sunburning, and dietary history; (b)
medical history, including use of phototoxic drugs, history of diabetes,
hypertension, use of steroids, aspirin, antihypertensive and cardiac medi-
cation, and tobacco use; and (c) exposure history, including a detailed
occupational hiistory covering each year of life from the age of 16 years, in
which daily exposure data was recorded on a monthly basis, and a similar
history for leisure activities. Both histories included information about
hours spent outside, hat use, glasses use, and history of arc welding.
Eye Examination
The visual acuity for each eye was tested with correction if the patient
wore distance glasses. If the acuity was less than 20/20, a pinhole test was
performed. If the vision was still less than 20/20, a subjective refraction
was done and the best corrected acuity determined. Intraocular pressure
was measured with either an applanation pneumotonometer or a hand-
 UV and the Eye 819

held applanation tonometer. Patients with elevated pressures were checked

with a Goldmann applanation tonometer.
Clinical Grading of Lens Opacities. The pupils were dilated with 10%
phenylephrine and 1% tropicamide. After dilation, lens opacities were
graded by slit-lamp examination using direct and retroillumination.115
The examination was considered adequate if the pupil was dilated to at
least 6 mm. Corneal and conjunctival abnormalities were also recorded.
The assessment of the lens status was made without the examiner know-
ing the patient's visual acuity.
Lens changes were clinically graded according to their anatomical
location and their estimated severity:
a) Cortical opacity-opacities seen on retroillumination that were
in the external 1/3 of the lens. These were graded for severity by
the estimated area of the lens affected by spokes. The area of the
opacities was expressed as the cumulative number of 1/8 wedges
of the retroilluminated lens involved with opacities, Grade 0 =
no opacity; grade 1 = < 1/8; grade 2 = < 1/4; and so forth.
b) Nuclear opacity-opacities seen in the optical section of the
inner 2/3 of the lens. These were graded according to their
expected impact on visual acuity as follows: grade 0 = no opac-
ity; grade 1 = some opacity but consistent with vision 20/20;
grade 2 = opacity consistent with vision between 20/20 and
20/30; grade 3 = opacity consistent with vision 20/40 to 20/100;
grade 4 = opacity consistent with vision 20/200 or less. Standard
photographs were used to assist in this grading. 115 The observer
was specifically not informed of the measured visual acuity to
minimize observer bias. This grading scheme was an important
innovation as it gave an unbiased assessment of nuclear opacity.
The bias of knowing the visual acuity has been a major disability
in many previous studies that have graded cataract.79-81 If the
examiner were aware of the subject's visual acuity, it could in-
duce observer bias, as the knowledge of an impairment could
encourage the examiner to look for and identify minimal opac-
ities and to consider them as significant.116
c) Posterior subcapsular cataracts (PSC). The overall height and
width of the opacity were measured with the slit lamp graticule.
d) Aphakia. Any lens that had been operated on, even if lens
remnants were present, was regarded as aphakia. There were 18
patients with aphakia. The operating ophthalmologist was con-
tacted in each case, and the preoperative diagnosis of the type of
cataract was obtained for 17 of them.
820 Taylor
We did not regard the following lens opacities and changes as senile
cataract: cortical vacuoles, Mittendorfs dots, small axial embryonal sul-
tural opacities, and other small dot opacities of a characteristic congenital
appearance (such as blue dot congenital opacities).
Photograding of Lens Opacities. All lenses were photographed in an
attempt to get a completely objective and unbiased assessment of lens
opacity. Retroillumination photographs of the anterior and posterior cor-
tex of the lens of each eye were taken using a Neitz cataract camera. 115 Slit
photographs of the nucleus were taken with a Topcon SL-5D photo slit
lamp using standard settings.17 One photographer took all the cortical
photographs and another took all the nuclear photographs.
Each set of photographs of the lens was graded for the presence and
severity of opacity. Cortical opacities were graded according to criteria
similar to those described for the clinical examination. Nuclear opacities
were graded by comparing them to the set of standard photographs of
nuclear opacities that corresponded to the clinical grading.'17
The photographs of the lens were graded independently by two grad-
ers. Any disagreement were adjudicated by a third grader. The photo-
graphs of the cortex were graded independently of the photographs of the
nucleus.
Analysis of the grading data showed good agreement between the
clinical grading and the photograding and good intra- and interobserver
agreement among graders."15,117,118
Photograding of Macular Changes. Subtle macular changes and early
stages of macular degeneration are often easier to detect on high-quality
fundus photographs than on clinical examination. Stereo fundus photo-
graphs centered on the disc and macula of each eye were therefore taken-
for each patient using a Topcon FET retinal camera.
Fundus photographs were graded with a new scheme we have devel-
oped to grade the presence and severity of macular changes."l9 This
grading scheme includes both the earliest fundus changes of macular
degeneration and drusen characteristics thought to be associated with an
increased risk of developing the exudative forms of this disease. In
examining the fundus photographs, the graders used a template with 3000
and 6000 pu circles and a set of standard photographs."19
Macular changes were graded as follows.
Grade 1: Drusen present within 3000 pu of the fovea (greater than
standard photo 1).
Grade 2: Drusen present within 1500 ,u of the fovea (greater than
standard photo 2).
 UV and the Eye 821

Grade 3: Presence of drusen within 1500 ,u of the fovea with so-called

"high-risk" characteristics (large, soft, confluent, pigment-
ed; these features were defined in standard photographs).
Grade 4: Presence of exudative disease or extensive geographic or
nongeographic atrophy (greater than a standard photo-
graph).
These categories were then combined as follows: SMD 1, grades 1
through 4; SMD 2, grades 2 through 4; SMD 3, grades 3 and 4; and SMD
4, grade 4 alone.
All fundus photographs were graded independently by two graders.
After each grader had read each photograph independently, any disagree-
ments were reviewed by the two graders together and the grade estab-
lished by consensus. The reliability and reproducibility of this scheme has
been assessed, and the interobserver and intraobserver agreement was
good to excellent."19
Skin Examination
We were interested in assessing the relationship of sun-induced skin
changes in both exposure and ocular damage. Each person had a der-
matologic examination. Particular attention was paid to the skin of the
head and neck. The degree of elastosis of the periorbital areas, including
the upper cheek and lower forehead, was graded in four steps from none
to severe according to a grading scheme we had devised.'20
Standard macrophotographs of the periorbital region were also taken
and later graded for actinic elastosis. Two graders compared skin photo-
graphs to standard photographs and graded each photograph by consen-
sus. The definitions were the same as those used in the clinical grading
scheme. Disagreements were adjudicated by a third dermatologist. Again,
inter- and intraobserver trials showed a high degree of reproducibility.120
The photograding is well correlated with the clinical grading of actinic
elastosis and the histologic grading of elastotic change.94
The examining dermatologist also recorded the presence of other skin
lesions, especially those due to excessive sun exposure, such as squamous
cell carcinoma, basal cell carcinoma, and actinic keratosis. When scars
were found (indicating the surgical removal or biopsy of lesions), the
treating physician was contacted and the diagnosis obtained from the
medical records in all cases.
Examination Procedures
The examinations were conducted in each community at a suitable local
building such as a church or fire hall. When a waterman came for the
examination, he underwent the following procedures: (1) his name was
822 Taylor

checked on a master list, informed consent was obtained, and a unique

identification number was assigned; (2) visual acuity and intraocular pres-
sure were measured, and dilating drops were instilled; (3) the interview
and skin examination with photography were performed while the dilat-
ing drops were taking effect; and (4) the eye examination was performed,
and lens and fundus photographs were taken. The ocular findings were
then discussed and any questions answered. If further follow-up or treat-
ment appeared warranted, appropriate referrals were made to local medi-
cal specialists.
Thirty-seven watermen were either unable or unwilling to come to the
central site; for them, a home visit was arranged. An interview was
obtained and an ophthalmologic examination undertaken. The lens was
examined with a portable slit lamp and, where possible, fundus photos
taken with a portable fundus camera. Skin photographs were also taken,
although the detailed skin examination was not done.
At each stage of data collection, information in each area was gathered
and recorded independently without knowledge of data from the other
areas in order to avoid bias. Thus, for example, interviews had no knowl-
edge of the status of the lens before they interviewed the subject on UV
exposure.
DATA ANALYSIS
Data were collected, for the most part, on self-coding forms. Data were
entered on personal computers using customized interactive data entry
programs (Infostar). Data files were transferred to a mainframe computer
(IBM 4331). The files were checked and edited. A number of standard
statistical packages were used during the analysis including SAS and
SPSS.
Calculation of Exposure
The yearly ocular UV-B exposure was calculated for each individual for
each year of life beyond the age of 15. This was done by combining
laboratory-derived and field-derived data and published data on ambient
UV-B with personal exposure histories obtained in interviews (Fig 1).
Laboratory investigations, using anatomically correct mannikin models,
were conducted to determine the ocular dose as a fraction of ambient UV-
B-the ocular ambient exposure ratio or OAER-and the attenuation of
ocular exposure by spectacles and hat use. 121 Seasonal variations in OAER
were examined during field studies in which watermen wore UV-B ab-
sorbing polysulfone film badges in order to monitor their actual daily
doses of UV-B while they worked on the water, crabbing or oystering,
 UV and the Eye 823
12
xp - E xPm [Fwork + Fi.is

m = month
eXPm mean ambient monthly exposure (MSY)
-
Fwork = fraction of monthly ocular exposure acquired at work
Fwork - Hrwork x LOcamb x Splcwork x OAERworks x (workdays)
7
Hrwork = fraction of daily ambient exposure during hours worked
Locamb = mean annual level at job location (compared to Maryland)
Specwork = mean UVR attenuation due to spectacle wear while working
OAERwork = OAER for given work surface, hat use, and season
Workdays = number of days worked per week

Fleis = fraction of monthly ocular exposure acquired during leisure

Fbis - Hr1*i5 x LOcamb x Spesqis x OAERi.is x (1 - workdays)
7
Hrleis = fraction of daily ambient exposure during leisure hours
Locamb = mean annual level at leisure location (compared to Maryland)
SPECleis = mean UVR attenuation due to spectacle wear on days off
OAERleis = OAER for given hat use and season
Workdays = number of days worked per week
FIGURE 1
The factors included in the mathematical model for the calculation of yearly ocular UV-B
exposure (Yexp) for each individual. (Yearly exposure is expressed as a fraction of total
ambient UV-B radiation in terms of Maryland Sun Years. See text for a more detailed
description of this model and the different factors.)

with and without hats.'22 Other outdoor workers were also studied to
obtain OAER for work over land. Data on the proportion of exposure for
each hour, for each month, and for different localities were available from
published tables.123124
The annual exposure was expressed as a proportion of total ambient
UV-B radiation in terms of Maryland Sun Years (MSY); one MSY is equal
to the total amount of ambient UV-B irradiance of a horizontal surface at
sea level in Maryland over 1 year. On the basis of readings taken in
824 Taylor
Philadelphia, this is approximately 2500 minimal erythemal doses, or 95J
cm -2.123,124 Once the individual yearly exposures were determined, it
was possible to determine other measures for each waterman, including
cumulative exposure, average annual exposure, maximum annual expo-
sure, age of first maximum exposure, number of consecutive or noncon-
secutive years at either his maximum annual exposure or a given level of
exposure.
Statistical Methods
A tremendous amount of data was collected during this study, and an
ordered approach to statistical analysis was followed. Some of the more
important statistical tests that were used are outlined.
The odds ratio is a test used to assess risk in retrospective or cross-
sectional studies. 125 It gives a general measure of the risk of disease for an
individual exposed to a specific factor in comparison with the risk of
disease for an individual not exposed. Specifically, it divides the odds that
an individual with the disease has been exposed to the risk factor by the
odds that an individual without the disease has been exposed to the risk
factor. An odds ratio of 1 means that there is no increased odds of having
been exposed to the risk factor if one has the disease. An odds ratio of 2
means that the odds of having been exposed to that factor are twice as
great in those with the disease. The 95% confidence interval (CI) is used
to assess whether an increased risk is statistically significant. If the CI of a
given odds ratio does not include 1.0 the increase in risk is statistically
significant, at least to the 0.05 level. for example, the increased odds of
cortical cataract if cumulative UV-B exposure is doubled is statistically
significant, as the derived odds ratio is 1.60 (95% CI, 1.01 to 2.64). The
odds ratio of 1.60 also indicates that people with the higher exposure have
60% more cataract than those with a lower exposure.
Data were analyzed using the SAS standard statistical package to calcu-
late Mantel-Haenszel summary odds ratios. Odds ratios were also derived
from logistic regression models used to analyze the independent contri-
bution of exposure to risk of the observed ocular changes.
Various regression analytic methods were also used to study the associ-
ation between one dependent variable (cortical opacities, for example)
and many independent variables (say age, UV-B exposure, smoking, and
hypertension). These sophisticated computer-based statistical techniques
can simultaneously indicate the strength and statistical significance of
multiple potential associations while controlling for possible confounding
effects of the independent variables. 125 Results from such an analysis can
often be presented as odds ratios but for continuous variables the regres-
sion coefficient with or without a confidence interval is usually presented.
 UV and the Eye 825
This coefficient gives a measure of the strength of the association for each
unit of that variable, for example, the regression coefficient for age in the
multiple logistic regression model of UV-B exposure and cortical opacities
is 0.16. This can be translated to say that, on average for each additional
year of life, the risk of cortical cataract increases by 17% (eO 16).
A serially additive expected dose model was also used to explore the
relationship between UV-B exposure and lens opacities. 126This method of
analysis involves calculating the actual yearly exposure for each year of life
for cases-for example, those with established cortical cataract-and com-
paring this with the exposure of age-matched controls who do not have
cortical cataract. Cases and controls were matched within 2 years of age.
RESULTS

REPRESENTATIVES OF STUDY POPULATION

The first step of the analysis was to ensure the representativeness of our
sample and show that it was not unduly affected by selection biases. The
random sample appeared to be representative of the sample of all nonat-
tenders on the two variables that could be determined from the DNR
records: age and primary license type (Table I).
TABLE I: COMPARISON OF AVAILABLE BASELINE CHARACTERISTICS OBTAINED FROM LICENSE
RECORDS (AGE AND LICENSE TYPE) OF PARTICIPANTS (STUDY POPULATION) AND THOSE WHO WERE
ELIGIBLE BUT DID NOT PARTICIPATE (NONATTENDERS)*
STUDY POPULATION
PRIMARY SECONDARY RANDOM
ATTENDERS ATTENDERS SAMPLEt TOTAL NONATTENDERS
NO * NO % NO % NO % NO %
Age (yrs)
30-39 145 23 47 26 12 29 204 24 103 27
4049 104 17 34 19 10 24 148 18 73 20
50-59 130 21 28 16 8 19 166 20 70 19
60-69 131 21 42 24 4 10 177 21 53 15
70-79 84 14 18 10 5 12 107 13 48 13
80+ 27 4 7 4 2 5 36 4 13 4
License
Crab 385 62 98 56 25 61 508 60 229 64
Oyster 174 28 46 26 8 20 228 27 81 22
Limited crab 12 2 15 9 3 7 30 4 17 5
Other 50 8 17 10 5 12 72 9 33 9
Total 621 176 41 838 360
*No statistically significant differences were seen between the groups.
Age: Random sample vs nonattenders: x2 = 0.95, P = NS.
License: random sample vs nonattenders: x2 = 1.04, P = NS.
tThe random sample subjects aged less than 30 or who only had interviews are excluded
from this table.
826 Taylor

The distribution of several variables between attenders and the random

sample was then examined to see if these groups were similar. Again there
were no differences with respect to age and license type (Table I). Fur-
ther, no difference was found in general medical background, such as the
tendency to freckle (present in 48%), frequency or degree of sun burning
(33% usually or always burn), or use of photosensitizing drugs (2%); nor
was there a difference in frequency of diabetes (7%), hypertension (34%),
reported use of heart or diuretic medication (20%), regular aspirin use
(36%), or use of oral steroids (4%). There was no difference in educational
history (average, 9.5 grades), smoking history (79% greater than 5 packs in
their lifetime), or the use of glasses outdoors (73%). Those in the random
sample had used arc welding more frequently than those in the other two
groups (49% compared to 26% and 27%). There was also no difference in
the hours spent outside in days off work (the median was 5 hours),
although 8% of the random sample did not wear a hat during leisure time,
compared to 21% and 20% of the primary and secondary attenders,
respectively.
Of particular concern would be a difference in the rates of cataract or
UV-B exposure. No difference was observed in either the proportion of
subjects with nuclear or cortical opacity (Table II) or several indices of UV-
B exposure, including average annual exposure, maximum annual expo-
sure, and age at first year of maximum exposure. The groups had similar
eye examinations. There was no difference in the frequency with which
corneal opacities occurred (85%), nor were there any differences between

TABLE II: DISTRIBUTION OF LENS OPACITES (CLINICAL EXAM) AMONG AITENDERS AND
RANDOM SAMPLE*
PRIMARY SECONDARY RANDOM
ATFENDERSt A1ITENDERSt SAMPLE
NO % NO % NO %

Nuclear opacity
None 462 75 116 66 25 63
1 88 14 34 19 8 20
2 or more 67 11 25 14 7 17
Cortical opacity
None 533 86 153 88 35 87
1 60 10 11 6 3 8
2 or more 24 4 11 6 2 5
*No statistically significant differences in the distribution of lens opacities are seen between
the groups.
Nuclear opacity: x2 = 7.26, P = NS
Cortical opacity: x2 = 3.71, P = NS
tTwo subjects with bilateral congenital cataract, three without clinical examinations, and one
with bilateral aphalda are excluded.
 UV and the Eye 827
the frequency of a history of glaucoma (2%) or the presence of pseudoex-
foliation (0.4%). Aphakia was equally common in each of these three
groups (3%).
These data suggest that, although not all of the baseline population was
examined, the attenders (who represent 70% of this population) and the
nonattenders were very similar on all the important characteristics that
could be assessed, and no bias from this source should be expected in the
results.
DESCRIPTION OF STUDY POPULATION

General and Occupational Characteristics

Of the 838 watermen examined, 204 (24%) were between 30 and 39 years
of age and 140 (17%) were over the age of 70 (Table I). The oldest was 94,
and the mean age was 53.0 years. There were 24 watermen (3%) who were
black.
During the interview, a detailed occupational history was obtained.
Overall, 75% had spent up to 4 years in military service, but approx-
imately 23% of the watermen had not held any other job apart from
military service. Less than 25% had held three or more jobs and less than
2% had held six or more. Altogether, information was collected on 1717
different jobs undertaken by members of the sample during their working
life. Of all these jobs, 72% were undertaken in Maryland and another 8%
in contiguous states.
Watermen could be classified according to their "main job." The main
job was defined as the occupation they had pursued for the longest
period, though not necessarily continuously. The various occupations
were grouped together into nine broad categories (Table III). Watermen
were regarded as working full time if they worked on the water year
TABLE III: CHARACTERISTICS OF OCCUPATIONS DEFINED AS MAIN JOBS
TIME JOB HELD HOURS OUTSIDE
NO (%) YEARS WORK DAYS LEISURE DAYS
Full-time watermen 437 (52.1) 28.2 ± 14.1 11.3 + 1.7 4.6 ± 2.7
Part-time watermen 103 (12.3) 25.6 ± 13.6 NA 5.1 ± 2.6
Water related 16 (1.9) 26.1 ± 15.1 7.5 ± 3.1 5.4 ± 2.2
Outside workers 80 (9.5) 22.7 + 12.0 5.3 ± 3.1 5.6 ± 2.3
Laborers 23 (2.7) 20.3 ± 10.8 2.3 ± 3.3 5.3 ± 2.3
Farmer 17 (2.0) 27.4 ± 12.5 8.6 ± 2.7 5.5 ± 3.0
Army 11 (1.3) 20.2 ± 5.0 3.0 ± 2.9 5.0 ± 3.3
Inside workers 108 (12.9) 22.2 ± 10.4 1.9 ± 2.7 5.1 ± 3.0
White collar 43 (5.1) 21.7 ± 9.9 2.0 ± 2.7 5.1 ± 2.8
Total 838 25.9 ± 13.3
828 Taylor
round. Part-time watermen would only work part of the year (the crab-
bing season, for example) and had either another job or no job for the rest
of the year. Water-related jobs included those of sailors and marine police.
Outside workers were those who spent most of their working hours
outside and included construction workers and truck drivers but not
farmers and laborers. The other categories are self-explanatory. The time
the main job had been held was similar for each category (Table III).
Although there was the expected variation in the number of working
hours spent outside, the number of leisure hours spent outside was
similar for each category.
The characteristics of the watermen's jobs were examined in detail;
overall, 705 (84%) of those studied had crabbed, 611 (73%) had oystered,
92 (11%) had fished, and 37 (4%) had clammed. Whatever occupation a
waterman was following, he was likely to be on the water from first light.
Those fishing, clamming, and crabbing usually returned during the mid-
dle of the day (12 noon to 2 pm), whereas those oystering usually did not
return until after 4 p.m.
The oystering season is set by regulation; it usually runs from October
to April. Most watermen will crab between April and September. Clam-
ming could start any time of the year, but most clammed for 2 to 4 months,
usually between April and June. Similarly, fishing could also be under-
taken year round, and 60% of watermen who fished did so for more than 6
months of the year. When crabbing, three-quarters would work 6 or 7
days a week, as would almost half of those fishing or clamming. Because of
the poor weather in winter, two-thirds would oyster only 4 or 5 days a
week.
Over three-quarters of watermen always wore a hat while working ,on
the water; less than 10% said they never wore a hat on the water. There
was no difference for the different types of watermen jobs. Oystering and
fishing are done without a canopy on the boat, although one-fifth of
clammers use a canopy and two-thirds of crabbers use a canopy at least
some of the time.
Distribution of UV-B Exposure
The annual exposure of the eyes to UV-B was determined for each
waterman for each year of his life after the age of 15 years. Given the
diverse occupational and behavioral background of the watermen, it is not
surprising that there was quite a range in the various parameters calcu-
lated. Cumulative exposure, for example, ranged from 0.024 to 3.664
MSY with the median being 0.756 MSY (Fig 2). Average annual exposure
ranged from 0.001 to 0.074 MSY with the median being 0.022 MSY (Fig
3). Maximum annual exposure ranged from 0.002 to 0.128 MSY with a
 UV and the Eye82 829
150
125
100
FREQUENCY 75
50
25
0 .1 .5 .9 1.3 1.7 2.1 2.5 2.9 3.3
3.7
CUMULATIVE EXPOSURE
FIGURE 2
Frequency distribution of cumulative ocular UV-B exposure of 838 watermen as a proportion
of total annual ambient UV-B (in Maryland Sun Year [MSY] units).

150
125 Median

100

FREQUENCY 75lii

25

.002 .012 .022 .032 .042 .052 .062 .072

AVERAGE EXPOSURE
FIGURE 3
Frequency distribution of average annual ocular UV-B exposure for 838 watermen as a
proportion of total annual ambient. UV-B (equal to 1 MSY).
830 Taylor
median of 0.038 MSY. For more than half of the watermen, their first year
of their maximum annual exposure had occurred by the age of 18 years.
The maximum possible annual ocular exposure in Maryland is 0.170 MSY.
To achieve this exposure, one would have to work on the water all day,
every day of the year, and not wear a hat or glasses. Under these circum-
stances, the eye would receive 17% of the ambient UV-B.
Distribution of Lens Opacities and Cataract
There was no difference in the occurrence of various types or severity of
senile lens opacities between the right eye and the left eye, and the
subsequent analysis has been performed for each type of opacity using the
grading in the worst or most severely affected eye.
Two watermen had bilateral congenital cataracts (one was aphakic) and
have been excluded from the subsequent analyses of lens opacities. Two
watermen had unilateral congenital lens opacities, and ten watermen had
unilateral traumatic opacities (two were aphakic). These 12 people were
classified according to the change in their unaffected eye. Seventeen were
aphakic, having had cataract surgery for senile cataract. Nine had uni-
lateral aphakia; and of these, four had moderate or severe cataract in the
fellow eye (Table IV).
Data on the clinical assessment of lens opacities was available on 835
watermen. Photographic data was available on cortical opacities for 749
and on nuclear opacities for 729. The reasons photographs were not
available are shown in Table V.
Overall, 111 (13%) watermen had some cortical opacity on clinical
examination and 229 (27%) had some nuclear opacity. These figures
include those who had had cataract surgery. There was a progressive

TABLE IV: DISTRIBUTION OF CATARACT TYPES IN APHAKIC PATIENTS*

UNILATERAL APHAKIA
FELLOW EYEt
BILATERAL OPERATED
APHAKIA EYE CO NO PSC

Cortical opacity (CO) 2 0

Nuclear opacity (NO) 4 3 0 2 0
Posterior subcapsular 1 5 0 0 2
(PSC)
NO and PSC 0 1 0 0 0
Unknown 1 0
Total 8 9
*One patient with unilateral aphakia following traumatic cataract is excluded.
tPresence of lens opacity grade 3 or greater.
 UV and the Eye 831

TABLE V: REASONS FOR MISSING OR UNREADABLE PHOTOGRAPHS

CORTICAL NUCLEAR MACULAR
PHOTOS PHOTOS PHOTOS
One eye
Aphakic 9 11 2
Monocular 4 4 4
Inadequate photograph 35 6 0
Opaque media 0 0 7
No known reason 4 9 2
Total 58 30 15
Both eyes
Aphakic 4 3 4
Drops contraindicated 11 14 10
Home visit 32 34 27
Refusal 7 8 9
Equipment failure 24 35 1
Subject missed photo 8 2 5
Inadequate photograph 0 0 5
No known reason 3 13 0
Total 89 109 61

increase in the prevalence and severity of cortical and nuclear opacities

with age. This was true with both the clinical grading and photograding
(Figs 4 to 7). The trends are remarkably similar for both types of assess-
ment, with the exception that the photograding of nuclear opacities
tended to grade opacities as being more severe than did the clinical
grading. Only 14 watermen were found to have PSC, including the 7 who
had cataract surgery for PSC (Fig 8), and these numbers are too few to
provide firm age-specific prevalence rates.
There was a highly significant statistical correlation between cataract
severity and visual acuity (Table VI). Overall, 83% of those eyes without
cortical opacities saw 20/20 or better compared to only 58% of those eyes
with cortical opacities. Similarly, 91% of eyes without nuclear opacities
saw 20/20 or better compared to only 46% of those eyes with nuclear
opacities. In this subanalysis, eyes with conditions-other than the partic-
ular cataract type-that might reduce best corrected acuity were ex-
cluded.
Distribution of Macular Changes
Macular photographic data was available for 777 watermen. The reasons
for missing photographs are shown in Table V.
Macular changes became increasingly common with increasing age.
The overall prevalence of macular changes of any severity (SMD 1)
increased with age from 17% in those less than 40 years of age to 50% in
832 Taylor

70
60
50
00L 40
PREVALENCE
30
20 .4 .

10 - *
30 40 50 60 70 80
AGE
FIGURE 4
Prevalence and severity of cortical opacities as determined by clinical grading in the more
severely affected eye for 838 watermen.

% 40
PREVALENCE 3
30/
20 / S.
0~~~~~~~~~~/**

30 40 50 60 70 80
AGE
FIGURE 5
Prevalence and severity of cortical opacities as determined by photograding in the more
severely affected eye for 746 watermen for whom cortical photographs were available.
 UV and the Eye 833

100

80

00L 60
PREVALENCE
40 1

20/

30 40 50 60 70 80
AGE
FIGURE 6
Prevalence and severity of nuclear opacities as determined by clinical grading in the more
severely affected eye for 838 watermen.

100

80

% 60
PREVALENCE
40

20

0 30 40 50 60 70 80
AGE
FIGURE 7
Prevalence and severity of nuclear opacities as determined by photograding in the more
severely affected eye for 726 watermen for whom nuclear photographs were available.
834 Taylor

7 6/177
3/3%
6 zoclinical psc
5 m aphakic psc
NUMBER 4 3/166 3/105
OF CASES 2/2% 30/a

I02i
2/204
10/
2

30-39 50-59 70-79

40-49 60-69 180
AGE GROUP
FIGURE 8
The occurrence of PSC as determined by clinical grading in the more severely affected eye
for 838 watermen.

those over age 80 (Fig 9). Grade 3 changes or worse (SMD 3) were present
in 97 (12%) cases. Ten (1%) watermen had frank macular degeneration;
that is, disciform degeneration or geographic atrophy (SMD 4).
Distribution of Other Ocular Abnormalities
Corneal opacities were seen in only 17 watermen (Table VII), but pteryg-
ium and pinguecula were common. Of particular interest was the fre-
quent occurrence of climatic droplet keratopathy (CDK); 162 watermen
showed some degree of this change. One waterman had bilateral Terrein's
corneal degeneration. Two watermen had miscellaneous conjunctival con-
ditions; conjunctival dysplasia was found in one and Bowen's disease in
another.
Less than 2% of eyes were found to have pressures of over 21 mm Hg,
and only four eyes had pressures of 30 mm Hg or greater.
Distribution of Skin Disease
Skin changes due to sun damage were common. Seven hundred four
(89%) of the 788 watermen who had skin examinations had moderate to
severe elastosis on clinical grading, and 540 (66%) of the 819 who had skin
 UV and the Eye 835

TABLE VI: CORRELATION BETWEEN SEVERITY OF LENS OPACI AND VISUAL

ACUITY IN INDIVIDUAL EYES*
VISUAL ACUITY
20/20 20/2.20/30 20/40 OR LESS
Cortical opacityt
Grade 0 1032 102 14
Grade 1 46 28 7
Grade 2 or more 7 3 1
Total no. of eyes 1240
Nuclear opacityf
Grade 0 975 80 14
Grade I 110 51 7
Grade 2 or more 20 32 31
Total no. of eyes 1320
*Eyes with aphkia; congenital, traumatic or PSC lens opacities; SMD 3 or
4; or other obvious ocular pathology that would reduce acuity were
excluded.
tEyes with grade 2 nuclear opacity or worse were also excluded (x2 =
86.4, P < 0.001. Fisher's exact test [opacity present/absent vs 20/20-20/25
or worse] P < 0.001.)
*Eyes with worse than grade 2 cortical opacity were also excluded (X2 =
413.8, P<0.001).

60
--- SMD I
50-
......
SMD 2

--i SMD 3
401- SMD 4

PREVALENCE 30
20 F . 00
PI
10
-

V
nI I 0

30 40 50 60 70 80
AGE
FIGURE 9
The prevalence and severity of macular degeneration in the more severely affect eye as
determined by examination of macular photographs in 777 watermen.
836 Taylor
TABLE VII: OCCURRENCE OF CORNEAL ABNORMALITIES IN
835 WATERMEN
CORNEA UNILATERAL BILATERAL

Axial opacity 3 (0.4%) 0 ( -)

Peripheral opacity 13 (1.6%) 1 (0.1%)
Pterygium 70 (8.3%) 70 (8.3%)
Pinguecula 41 (4.9%) 601 (71.9%)
CDK 18 (2.1%) 144 (17.2%)

photography had moderate to severe elastosis on photograding. However,

246 showed periorbital sparing due to the wearing of either glasses, a hat,
or both. Actinic elastosis, graded either clinically or photographically,
showed a progressive increase with age, although there was a tendency
for some plateauing after the sixth decade (Fig 10). The 24 black water-
men were excluded for this comparison.
Actinic elastotic change is assumed to be a function of UV-induced
damage and therefore may provide a measure of cumulative personal UV
exposure. When the protective influence of glasses and hats is excluded
from the calculations of cumulative exposure, exposure is well correlated
with degree of actinic elastosis (Fig 11). Whether the actinic elastosis is
graded clinically or photographically (data not shown), an increase in
mean cumulative exposure is seen, with more severe grades of elastosis
for most decades.

RISK FACTOR ANALYSIS

UV-B Exposure and Cataract
Cataracts are increasingly more common at older ages (Figs 4 to 7) and any
analysis of association or risk factors must control for the effect of age.
Furthermore, because cumulative UV-B exposure and age are so closely
related, particular care must also be taken when assessing the effect of
cumulative exposure. Because of the complexity of the following analysis
and because similar results were obtained using both the clinical and
photographic grading schemes, only the data for clinical grading will be
presented.
To identify potential confounding factors, age-adjusted bivariate anal-
yses were performed (Table VIII). Subsequent logistic regression analysis
adjusting for age revealed a statistically significant association between
total cumulative UV-B exposure and cortical opacities (regression coeffi-
cient = 0.70; 95% CI, 0.01 to 1.40) (Table IX). These data show that, ie,
for a given age, a doubling of cumulative UV-B exposure was associated
 UV and the Eye 837

100
t00o- -tmtwwoSo*-ff W
.4.
801-
0VL 60sF *

-
Grade
*0.

PREVALENCE .-- 1 or worse II

401- : --- 2 or worse II

...... 3 or worse 0I

mm 4

20!F .0.0 _ _. 4

_._ * * I I
0
30 40 50 60 70 80
FIGURE 10
AGE
The prevalence and severity of actinic elastosis determined by clinical assessment in 704
watermen. This analysis excludes 24 black watermen.
3.0
E Mild
2.5 0 Moderate
M Severe
2.0
MEAN
EXPOSURE 1.5
1.0

.5

0
30-39 50-59 70-79
40-49 60-69 P80
AGE GROUP
FIGURE 11
The correlation between the severity of clinically determined actinic elastosis and mean
cumulative skin UV-B exposure in 704 watermen. This analysis excludes 24 black watermen.
There were no watermen aged 70 to 79 years with mild actinic elastosis.
838 Taylor

TABLE VIII: AGE ADJUSTED ODDS RATIOS (WITH 95% CI) SHOWING THE ASSOCIATIONS BETWEEN
VARIOUS POTENTIAL RISK FACTORS AND DIFFERENT TYPES OF LENS OPACMES (CLINICAL GRADING)*
ODDS RATIO
CORTICAL NUCLEAR PSC
FACTOR NOt (n = 11) (n = 229) (n = 14)
Heart/diuretic medication 163 1.5 (0.9-2.4) 0.9 (0.4-1.4) 2.4 (0.8-7.2)
History of hypertension 286 0.9 (0.6-1.4) 0.6 (0.4-0.9)* 1.9 (0.7-5.6)
Smoking 660 1.3 (0.8-2.3) 1.6 (0.9-2.9) 0.9 (0.2-3.3)
Freckling 402 0.9 (0.5-1.4) 0.8 (0.5-1.3) 1.7 (0.5-5.6)
Steroid use 33 0.5 (0.1-2.1) 1.0 (0.4-2.4) 1.5 (0.2-10.8)
Tendency to burn 279 1.1 (0.7-1.8) 1.6 (1.0-2.4) 2.4 (0.8-7.2)
Aspirin use 301 1.1 (0.7-1.8) 0.9 (0.6-1.4) 0.4 (0.1-1.5)
Diabetes 62 1.3 (0.6-2.6) 1.0 (0.5-1.9) 0.7 (0.1-5.6)
More than 8 years school 502 1.3 (0.8-2.1) 0.8 (0.5-1.2) 1.0 (0.3-3.0)
Arc welding 227 1.0 (0.5-1.8) 0.7 (1.4-1.1) 1.4 (0.4-4.7)
Eye color (blue eyes) 480 0.6 (0.4-1.0)t 0.8 (0.5-1.3) 9.0 (1.5-54.0)f:
Actinic elastosis 422 0.8 (0.5-1.3) 1.1 (0.7-1.9) 1.3 (0.44.5)
Pterygium 140 0.8 (0.5-1.4) 0.8 (0.4-1.5) 3.3 (1.1-9.9)t
Pinguecula 640 0.8 (0.4-1.3) 1.0 (0.7-1.4) 0.3 (0.1-0.9)t
Glaucoma 15 0.5 (0.1-1.9) 0.6 (0.2-1.9) -§
*See Methods section, Data Analysis for a description of odds ratio and its significance.
tNumber of watermen having this condition.
4Chi-square test gave, P < 0.05.
§None of the PSC cases had glaucoma.

TABLE IX: RESULTS OF LOGISTIC REGRESSION ANALYSES OF THE ASSOCIATION BETWEEN

CUMULATIVE UV-B EXPOSURE AND RISK OF LENS OPACITY (GRADE 2 OR MORE)*
REGRESSION STANDARD
VARIABLE COEFFICIENT ERROR P

Cortical opacity
Intercept - 14.23 1.93 0.001
Age 0.16 0.02 0.001
Cumulative UV-B exposuret 0.70 0.35 0.04
Nuclear opacity
Intercept -10.88 0.81 0.0001
Age 0.17 0.01 0.001
Cumulative UV-B exposuret -0.01 0.16 NS
History of hypertension -0.53 0.22 0.02
*Statistically significant association seen between cumulative UV-B exposure and cortical
opacity but not nuclear opacity.
tLog transformation.
 UV and the Eye 839
TABLE X: ODDS RATIO (WITH 95% CI) DETERMINED BY LOGISTIC REGRESSION
ANALYSIS FOR CORTICAL AND NUCLEAR CATARACT (EACH GRADE 2 OR MORE) BY
QUARTILES OF ANNUAL AVERAGE UV-B EXPOSURE*
ODDS RATIOS
EXPOSURE DOSE RANGE
QUARTILE (MSY) CORTICAL NUCLEAR
1st 0.012 or less 1.00 1.00 (0.51-2.91)
2nd 0.013-0.022 2.32 (0.70-7.73) 1.22 (0.51-0.91)
3rd 0.023-0.032 3.25 (0.98-10.80) 1.26 (0.52-3.06)
4th 0.033 or more 3.30 (0.90-9.97) 0.96 (0.36-2.60)
Age (regression coefficient) 0.16 0.17
*This analysis shows an increased risk of cortical opacity but not
nuclear opacity with increased annual UV-B exposure.

with a 1.60 times (95% CI, 1.01 to 2.64) increased risk of cortical opaci-
ties. Examined in another way, the risk of cortical cataract increased
markedly with increasing annual average UV-B exposure (Table X). The
odds ratio for exposure in the upper quartile compared to the lowest
quartile was 3.30 (95% CI, 0.90 to 9.97). No increased risk for nuclear
opacities was apparent with increased UV-B exposure, even adjusting for
age, tendency to burn, and history of hypertension (regression coefficient
= -0.01; 95% CI, -0.32 to 0.31).
The UV-B exposure for cases of cortical opacity was significantly higher
than the expected exposure when examined in the serially additive dose
model (t-test, 2.23; P = 0.03; Fig 12). It is clear that those with cortical
opacities have had more UV exposure every year of life after age 15,
suggesting that damage is a cumulative phenomenon. On average, this
model indicates that people with cortical cataract had a 21% higher UV-B
exposure than those without cortical cataract. There is no evidence that
those with cortical opacities had more exposure at a particular susceptible
period of life or that there was an obligatory latency period for develop-
ment of this change. Further analysis found no evidence to suggest a safe
threshold for either exposure level (annual average UV-B exposure) or
duration (the number of years above a given UV-B exposure). Instead, the
data indicate a progressive increase in risk of cortical cataract with in-
creased UV-B exposure. No such association was seen between UV-B ex-
posure and nuclear opacities in the serially additive dose model (Fig 13).
Other Risk Factors for Cataract
The association between various other factors and lens opacities deter-
mined clinically is shown in Table VIII. Similar results were obtained
when photographic grading was used. The prevalence rates of cortical
opacities were higher in those with blue eyes. The prevalence rates for
840 Taylor
11
9
7
AVERAGE 5
DIFFERENCE 3
(MSYX103 ) 1 -
-1
-3
20 30 40 50 60 70 80
AGE
FIGURE 12
The average difference in ocular UV-B exposure at each year of life between 34 watermen
with cortical opacities (grade 2 or worse in the more severely affected eye) and 213 age-
matched controls without this degree of cortical opacity as determined by the serially
additive expected dose model. (Note: in this model, if there were no difference in exposure
between the cases and the controls, the line would hover around 0. If the cases had a lower
exposure, the line would be below 0. Here the line is above 0, indicating that those with
cortical cataract have a consistently higher exposure at every year of life above age 15. The
annual exposure of cases with cortical opacities is, on average, 21% higher than controls.)

9
7
AVERAGE 5
DIFFERENCE K
(MSYx103) 1
-1
-3
-5
20 30 40 50 60 70 80
AGE
FIGURE 13
The average difference in ocular UV-B exposure at each year of life between 184 watermen
with nuclear opacities (grade 2 or worse in the more severely affected eye) and 413 age-
matched controls without this degree of nuclear opacity as determined by the serially
additive expected dose model. (There is no difference in UV-B exposure between the cases
with nuclear opacities and controls, as the line hovers around 0.)
 UV and the Eye 841

nuclear opacity were somewhat lower in those reporting a history of

hypertension. However, because of the large number of subjects report-
ing uncertainty over their use of antihypertensive medication and be-
cause blood pressure was not actually measured, this finding should be
interpreted with caution. The prevalence of both cortical and nuclear
opacities was higher among those who had "ever smoked" than among
those who had "never smoked," although the difference was not signifi-
cant in the initial analyses. There was no correlation between the pres-
ence of either basal cell or squamous cell carcinoma and cortical or
nuclear opacities.
The number of cases of PSC was small, so the analyses for this opacity
were limited. The prevalence rates of PSC were significantly higher in
blue-eyed individuals. They were also significantly higher among those
with severe actinic elastosis as determined by photograding and among
those with pterygium (compared to those with pinguecula alone or nei-
ther disease). The rates of PSC were higher in steroid users and those who
used heart or diuretic medication, but these did not reach statistical
significance.
In contrast to recent reports suggesting a protective effect of aspirin in
cataract prevention, we found no evidence that aspirin use, duration of
use, or dose had any impact on opacities.127
UV-B Exposure and Senile Macular Degeneration
No correlation was found between the presence of SMD of any severity
and UV-B exposure as expressed by cumulative or annual average expo-
sure or by the other measures of UV-B exposure that were outlined in the
analysis performed for cataract (Table XI and Fig 14). Similarly, no associa-
tion was found between SMD of any severity and cortical or PSC lens
opacities, pterygium, or eye color. The presence of nuclear opacities was
associated with an increased risk of SMD grade 3, and unexpectedly,
smoking appeared to be somewhat protective for SMD. Analysis showed
no consistent association between SMD of any severity and any of the
other variables collected in the medical history, occupational history, or
examination.
UV-B Exposure and Corneal Disease
Both pterygium and CDK were more common in those who had a higher
UV-B exposure. This was true for both annual average UV-B exposure
(Table XII) and for cumulative exposure (data not shown). No such associa-
tion was seen with pinguecula. Because a previous or coexistent pin-
guecula is often overlooked in the presence of a pterygium, a further
842 Taylor
TABLE XI: ODDS RATIO (WITH 95% CI) AS DETERMINED BY
LOGISTIC REGRESSION ANALYSIS FOR MACULAR
DEGENERATION (SMD 3) BY QUARTILES OF AVERAGE
ANNUAL UV-B EXPOSURE*
EXPOSURE
QUARTILE DOSE RANGE (MSY) SMD 3

1st 0.012 or less 1.00

2nd 0.013-0.022 1.85 (0.96-3.53)
3rd 0.023-0.032 1.34 (0.67-2.67)
4th 0.033 or more 1.10 (0.53-2.28)
Age (regression coefficient) 0.45
*This analysis shows no increased risk of macular
degeneration with increased annual UV-B exposure.

11
9
7

AVERAGE 5
DIFFERENCE 3
(MSYx1O-3) I

-1
-3
-5

AGE
FIGURE 14
The average difference in ocular UV-B exposure at each year of life between 89 watermen
with macular degeneration (SMD 3 or worse in the more severely affected eye) and 670 age-
matched controls without this degree of macular degeneration as determined by the serially
additive expected dose model. (There is no difference in UV-B exposure between the cases
with SMD 3 and controls, as the line hovers around 0.)

TABLE XII: ODDS RATIO (95% CI) AS DETERMINED BY LOGISTIC REGRESSION ANALYSIS FOR
PTERYGIUM, PINGUECULA, AND CDK BY QUARTILES OF AVERAGE ANNUAL UV-B EXPOSURE*
EXPOSURE
QUARTILE DOSE RANGE (MSY) PTERYGIUM PINGUECULA CDK

1st 0.012 or less 1.00 1.00 1.00

2nd 0.013-0.022 1.01 (0.55-1.83) 0.98 (0.63-1.53) 1.78 (0.94-3.35)
3rd 0.023-0.032 1.93 (1.09-3.39) 1.17 (0.75-1.85) 3.42 (1.85-6.31)
4th 0.033 or more 3.06 (1.77-5.31) 1.40 (0.88-2.23) 6.36 (3.46-11.68)
Age (regression coefficient) 0.05 0.01 0.08
*This analysis shows the increasedd risk of pterygium and CDK but not pinguecula with
increased annual UV-B exposure.
 UV and the Eye 843

analysis was undertaken to detect a possible interaction between pin-

guecula and pterygium. UV-B exposure continued to show a strong
association with pterygium but not with pinguecula. Age-adjusted odds
ratios were calculated for the risk of developing each of these conditions
for each quartile of annual average UV-B exposure. A strong association
was seen for both pterygium and CDK but not for pinguecula (Table XII).

DISCUSSION
In this study, we set out to assess whether there was an association
between high UV-B exposure and different types of senile cataract or
SMD. To be able to do this, we had to develop methods for quantifying
personal UV exposure for each individual in our study. We had to develop
reproducible grading methods for assessing cortical and nuclear cataract
and SMD. In the final analysis, this study shows a clear association
between a high UV-B exposure and an increased risk of visually significant
cortical cataract (ie, one quarter or more of the cortex involved with
opacity). Those with this degree of cortical cataract had 21% greater
exposure to UV-B radiation at each year of life after the age of 15 years
than those who did not have cortical cataract; and a doubling of cumula-
tive UV-B exposure increased the risk of cortical cataract by 60%. No
association was found between UV-B exposure and either nuclear opaci-
ties or SMD.
Previous evidence from biochemical, animal, and epidemiologic stud-
ies suggesting that UV-B exposure might be associated with the occur-
rence of so-called senile cataracts in man was largely circumstantial, as
was weaker evidence to suggest some linkage with SMD. There was a
need for a definitive study relating the actual ocular UV-B exposure of
individuals with their ocular status.
We selected the watermen of Chesapeake Bay as our study population
for a number of reasons. First, and most obviously, watermen were
selected because they work outside all day. Potentially, they are exposed
to the maximum UV-B irradiation naturally possible in this region. They
usually work on an open boat; and because the land around the Chesa-
peake Bay is flat, they have an unobstructed horizon. Therefore, they are
exposed to scattered UV-B from the whole sky. Further, as they work over
water, they are also exposed to some additional reflected UV-B.
Second, although some watermen may be exposed to the maximum
amount of UV-B, there is a range of exposure within this otherwise fairly
homogeneous and stable population group. Many watermen wear hats,
some wear glasses, and others have held other occupations with varying
844 Taylor
UV-B exposure. These factors resulted in a wide range of individual UV-B
exposures within this group, which greatly facilitated our analysis.
The third reason for selecting the watermen is perhaps the most
important. Their work habits are, by and large, very stable and predict-
able. This is due, in part, to the nature of their work and, in part, to state
regulations. Most watermen are on the water before dawn, and the time
they return to the dock is often set by regulation. Seasons, catches, and
equipment are all regulated. The fact that the Chesapeake Bay skipjack is
the last working sailboat in the country attests to this fact. 128 The working
habits of a waterman today have changed very little from those of 10, 20,
or even 50 years ago. This is true for very few other outdoor occupations;
one has only to think of the changes that have occurred in farming over
the last 50 years with the use of tractors and then protective cabs.
All of these factors taken together mean that the watermen of the
Chesapeake Bay form a stable occupational group for whom one could
reasonably hope to determine work habits and, therefore, occupational
UV-B exposure over their working lives with an acceptable degree of
accuracy and who would have a broad range of personal UV-B exposure.
Before undertaking this study, we undertook sample size calculations to
ensure that we would examine sufficient numbers of watermen to be able
to answer the questions we were asking. Calculations repeated after the
data was collected showed the initial calculations were correct. For nucle-
ar opacities, which occurred most commonly (prevalence of 28%), we
could be confident of identifying factors that increased the risk of cataract
by as little as 40% (relative risk, 1.40) if they affected 30% of the popula-
tion (type 1 error, 0.05; type II error, 0.20). This would include factors
such as sunburning (33%), hypertension (34%), regular aspirin use (30%),
and those who had smoked less than five packs of cigarettes (21%). lor
factors that affected only about 5% of the population (such as diabetes
[7%], steroid use [4%r, and glaucoma [2%]), we could detect an 80%
increase (relative risk, 1.80). For less common changes such as cortical
opacities (prevalence of 13%) and macular changes (SMD 3, prevalence of
12%), we could be confident of detecting an increase of approximately
65% (relative risk, 1.65) for a factor that affected 30% of the population or
140% (relative risk, 2.40) for a factor that affected 5%.
Our sample size was therefore large enough to address the major
questions we posed on the effect of UV-B exposure on the occurrence of
cortical and nuclear cataract and SMD. The sample size was not sufficient
for detailed analysis of less common outcomes-in particular, PSC, of
which only 14 cases were seen. Similarly, the study was not designed to
have the power to test the association between cataract and relatively
 UV and the Eye 845

uncommon factors such as diabetes.

The calculation of individual ocular UV-B exposure is unique and goes
far beyond any previous attempt to quantify personal exposure to UVR.129
Innovative measurements and approaches allowed the development of a
set of values that quantified the way personal habits or characteristics
altered UV-B exposure. 11,12.121,122 The relatively stable work habits of the
watermen permitted the collection of detailed lifetime occupational infor-
mation. The integration of these two data sets in the ocular exposure
model permitted us to develop an assessment of individual ocular expo-
sure to UV-B with much greater precision than previously possible. The
correlation between the calculated dermal UV-B exposure and actinic
elastosis attests to the validity of our model, as does the distribution and
statistical behavior of the exposures themselves.
One of the major stumbling blocks in any study is the lack of a clear
definition of outcome or "a case." This was a major problem facing us at
the start of this study. We eventually had to develop grading schemes for
seven separate outcomes: clinical grading schemes for cortical, nuclear,
and posterior subcapsular opacities and photographic grading schemes for
cortical and nuclear opacities, macular changes, and actinic elasto-
sis. 115,117-120 In each instance, the grading scale gave a three- or four-step
gradation from no disease to severe disease. The progression of severity
and the proportion of disease at each grade is a good measure of the
soundness or biological plausibility of a scheme. As can be seen from our
data, there is good distribution of people in each grade and steady
progression of severity with increasing age for each grading scheme (Figs
4 to 9), with the sole exception of PSC for which too few cases were seen.
Another important measure of the soundness of a grading scheme is its
reproducibility. We have also demonstrated high levels of reproducibility
by both inter- and intraobserver reliability testing for each scheme. 115,117-120
On the one hand, the grading of photographs would seem to have much
to recommend it over clinical grading. The grading of photographs can
more easily be standardized and would seem to be less subject to observer
bias. Photograding readily lends itself to the masking of the observer;
multiple observers can be used and observer reproducibility trials can be
performed easily. While these benefits may seem commendable, we
found in this study that these apparent advantages come at some cost.
Overall, for one in eight of the watermen, cortical or nuclear photographs
were not obtained, most commonly because of equipment failure or
inability to transport the cameras on home visits. On the other hand, the
parallel analysis of the clinical data and the photographic data gave similar
results. This suggests that for future cross-sectioned field studies of cata-
846 Taylor

ract, the standardized clinical grading methods may be not only sufficient
but also preferable.
As discussed in detail in the Introduction, much of the biochemical
literature relating to UV-B exposure and cataract has focused on the
changes that occur in the nuclear pigmentation; and field studies have
often reported the occurrence of mature cataract. Taken together, they
have fostered the notion that the association between UV-B and cataract
should relate to nuclear and not cortical cataract. However, the biochemi-
cal evidence for UV-induced changes in the proteins of the cortex seems
to be at least as compelling as for changes in nuclear proteins. The
experimental data from animal studies, also, all point toward UV-B caus-
ing cortical changes and not nuclear changes. Furthermore, the three
epidemiologic studies that have separated cataract by type showed an
association between UV-B exposure and cortical cataract, not nuclear
cataract. 81,93,94 In this context, the finding of our study is quite consistent
and not unexpected.
The lack of an association between UV-B exposure and SMD is also not
altogether unexpected in that the lens absorbs most of the UV-B, and only
very small amounts of this waveband can reach the retina. It would be of
considerable interest to investigate the association between UV-A expo-
sure and SMD, since more UV-A is transmitted by the lens, and this
waveband also has the potential for retinal toxicity.
It is of interest to note the relative rarity of PSC changes in our
population-based sample. PSC occurred in only 14 watermen whereas
cortical opacities appeared in 111 and nuclear opacities in 229. However,
PSC had led to cataract surgery in 7 watermen (associated with nuclear
opacities in 1), while only 10 had cataract surgery as a result ofother types
of cataract. The watermen with PSC were also younger than those with
other lens opacities. This suggests that PSC, although relatively uncom-
mon, occurs in younger patients, develops more rapidly, affects vision
earlier, and therefore leads to earlier surgery than other lens opacities.
What is the clinical relevance of the strong epidemiologic and statistical
association we found between ocular exposure to UV-B and cortical cata-
ract? At present, we do not understand the causes of senile cataract in
general or cortical cataract in particular, even though over 1 million
cataract operations are performed each year,108 and approximately one
third of them are for cortical cataract. The cause of cataract now appears to
be clearly multifactorial. This study has for the first time convincingly
identified one of the important risk factors for cortical cataract. The
association between UV-B exposure and cortical cataract was strongest for
those with at least grade 2 cortical opacities-that is, where one-quarter
 UV and the Eye 847

or more of the cortex was already opacified. This degree of cortical

opacity was frequently associated with 20/30 vision or less. By clinical
experience, such opacities proceed quite rapidly to cataract surgery in the
majority of cases.
Given these findings, what are the implications for the watermen or for
the general population? It would seem prudent to protect the eyes from
unnecessary exposure to UV-B. The amount of ambient UV-B varies
markedly during the day (being highest in summer between 10 am and 2
pm). The periods of high levels of UV-B are usually well recognized, as
this is the time when one is most likely to become sunburned. As a public
health recommendation, therefore, people should be advised to use ocu-
lar protection at those times when they are at risk of getting sunburned.
There are two very easy ways to reduce UV-B exposure short of staying
indoors, out of the sun. A hat with a brim will reduce ocular exposure by
half, and ordinary, close-fitting plastic spectacles can reduce it to about
5%.12 The effect of hats and glasses are additive. Although special UV-
absorbing lenses can stop all UV-B transmission, a sample of 40 clear
spectacle lenses showed that all ofthem significantly reduced ocular UV-B
exposure.12 The shape and position of the glasses were more important in
reducing UV-B exposure than the actual lens material. Wrap-around
glasses gave almost complete protection, while, if the frames were one
inch from the brow, there was only a 25% reduction of the amount of UV-
B which reached the eye. Thus, to minimize ocular exposure to UV-B
cheaply and effectively, people should be advised to wear a hat with a
brim and close fitting sunglasses at times when they could get sunburned.

SUMMARY
Circumstantial evidence from biochemical, animal, and epidemiologic
studies suggests an association between exposure to UV-B radiation (290
nm to 320 nm) and cataract. Such an association had not been proven
because it had not been possible to quantify ocular UV-B exposure of
individuals or to reliably grade the type and severity of cataract in field
studies. We undertook an epidemiologic survey of cataract among 838
watermen who work on the Chesapeake Bay. Their individual ocular UV-
B exposure was quantified for each year of life over the age of 16, on the
basis of a detailed occupational history combined with laboratory and field
measurements of ocular UV-B exposure. Cataracts were graded by both
type and severity through clinical and photographic means. SMD changes
were ascertained by fundal photography. A general medical history was
taken to discover potentially confounding factors. This study showed that
848 Taylor
people with cortical lens opacities had a 21% higher UV-B exposure at
each year of life than people without these opacities. A doubling in
lifetime UV-B exposure led to a 60% increase in the risk of cortical
cataract, and those with a high annual UV-B exposure increased their risk
of cortical cataract over threefold. Corneal changes, namely pterygium
and CDK, were also strongly associated with high UV-B exposure. No
association was found between nuclear lens opacities or macular degener-
ation and UV-B exposure. This study also indicated several simple, practi-
cal measures, such as wearing spectacles or a hat, that effectively protect
the eye from UV-B exposure. Thus it is easily within the power of
individuals to protect their eyes from excessive UV-B exposure and re-
duce their risk of cortical cataract. A program of public education in this
area could be a cost-effective means of reducing this important disease.

ACKNOWLEDGMENTS

The author wishes to thank the many people who assisted with this
arduous and exacting undertaking. Special thanks are due to Drs Sheila
West, Edward Emmett, Helen Abbey, Frank Rosenthal, Stuart Fine, Neil
Bressler, Susan Bressler, Lorraine Cameron, and Mrs Beatiz Munoz. Drs
Henry Newland, Ben Vitasa, and Alex Bakalian assisted with field work as
did Miss Paula Prestia and Messrs David Emmett and John Scrimgeour.
Invaluable help and assistance came from members of the Maryland
Waterman's Association, especially Mr Wayne "Hon" Lawson of Crisfield.
The dedicated assistance of Miss Sue Simmons, Mrs Celeste Wilson, and
Mrs Alice Flumbaun is acknowledged as well as the editorial assistance of
Dr Susan Edmunds and the support and advice of Dr Elizabeth Dax.
REFERENCES
1. Guerry RK, Ham WT, Mueller HA: Light toxicity in the posterior segment, in TD
Duane, EA Jaeger (eds): Duane's Clinical Ophthalmology, Vol 3. Philadelphia, Harper
& Row, 1986, pp 1-17.
2. American Academy of Ophthalmology: Eye and visual system abnormalities, in Eye
Care for the American People, Supplement to Ophthalmology. San Francisco, Ameri-
can Academy of Ophthalmology, 1987, Chap 2.
3. Hausser KW, Vahle W (translated by F Urbach): Sunburn and suntanning, in F Urbach
(ed): Biologic Effects of Ultraviolet Radiation. Oxford, Pergamon Press, 1969, pp 3-21.
4. Jones RR: Ozone depletion and cancer risk. Radiat Res 1987; 2:443-446.
5. Duke-Elder S, MacFaul PA: Radiational injuries: Action on the lens, in S Duke-Elder
(ed): System of Ophthalmology. St Louis, CV Mosby, 1972, Vol XIV, Pt 2, pp 928-933.
6. Phillips M: Electromagnetic radiation, in Encyclopaedia Brittanica, Vol 6. Chicago,
Encyclopaedia Brittanica, Inc, 1980, pp 644-665.
7. Parrish JA, Anderson RR, Urbach F, et al: UV-A: Biological Effects of Ultraviolet
Radiation with Emphasis on Human Responses to Longwave Ultraviolet. New York,
Plenum Press, 1978, Chap 1.
 UV and the Eye 849
8. Robertson DF: Solar ultraviolet radiation in relation to sunburn and skin cancer. MedJ
Aust 1968; 2:1123-1132.
9. Fishman GA: Ocular phototoxicity: Guidelines for selecting sunglasses. Surv Ophthal-
mol 1986; 31:119-124.
10. Sliney DH: Physical factors in cataractogenesis: Ambient ultraviolet radiation and
temperature. Invest Ophthalmol Vis Sci 1986; 27:781-90.
11. Rosenthal FS, Safran M, Taylor HR: The ocular dose of ultraviolet radiation from
sunlight exposure. Photochem Photobiol 1985; 42:163-171.
12. Rosenthal FS, Bakalian AE, Taylor HR: The effect of prescription eyewear on ocular
exposure to ultraviolet radiation. Am J Public Health 1986; 76:1216-1220.
13. Sliney DH: Eye protective techniques for bright light. Ophthalmology 1983; 90:937-
944.
14. Coroneo MT: Albedo concentration in the anterior eye and the location of some solar
diseases. MS thesis. Sydney, Australia, University of Sydney, 1988.
15. Kinsey VE: Spectral transmission of the eye to ultraviolet radiations. Arch Ophthalmol
1948; 39:508-513.
16. Bachem A: Ophthalmic ultraviolet action spectrum. Am J Ophthalmol 1956; 41:969-
975.
17. Boettner EA, Wolter JR: Transmission of the ocular media. Invest Ophthalmol Vis Sci
1962; 1:776-783.
18. Cooper G, Robson J: The yellow colour of the lens of man and other primates. J Physiol
(Lond) 1969; 203:411-417.
19. Lerman S: Human ultraviolet radiation cataracts. Ophthalmic Res 1980; 12:303-314.
20. Rosen ES: Filtration of non-ionizing radiation by the ocular media, in J Cronly-Dillon,
ES Rosen, J Marshall (eds): Hazards of Light: Myths and Realities of Eye and Skin.
Oxford, Pergamon Press, 1986, pp 145-152.
21. Zigman S: Photobiology of the lens, in H Maisel (ed): The Ocular Lens: Structure,
Function, and Pathobiology. New York, Marcel Dekker, 1985, pp 301-347.
22. VerhoeffFH, Bell L, Walker CB: The pathological effects of radiant energy on the eye:
An experimental investigation with a systematic review on the literature. Proc Am
Acad Arts Sci 1916; 51:630-818.
23. Pitts DG: The human ultraviolet action spectrum. Am J Optom Physiol Optics 1974;
51:946-960.
24. Pitts DG, Cullen AP, Parr WH: Ocular ultraviolet effects from 295nm to 335nm in the
rabbit eye. Washington, DC, U.S. Department of Health, Education, and Welfare
(NIOSH), 1976, pp 1-55.
25. Parrish JA, Anderson RR, Urbach F, et al: UV-A: Biological Effects of Ultraviolet
Radiation with Emphasis on Human Responses to Longwave Ultraviolet. New York,
Plenum Press, 1978, Chap 9.
26. Ham WT Jr, Mueller HA, Ruffolo JJ Jr, et al: Action spectrum for retinal injury from
near-ultraviolet radiation in aphakic monkey. Am J Ophthalmol 1982; 93:299-306.
27. Ham WT Jr, Ruffolo JJ Jr, Mueller HA, et al: Histologic analysis of photochemical
lesions produced in rhesus retina by short wavelength light. Invest Ophthalmol Vis Sci
1978; 17:1029-1035.
28. Spector A: The search for a solution to senile cataracts (Proctor Lecture). Invest
Ophthalmol Vis Sci 1984; 25:130-146.
29. Garner WH, Spector A: A preliminary study of the dynamic aspects of the age
dependent changes in the abundances of human lens polypeptides. Doc Ophthalmol
1979; 18:91-99.
30. : Racemization in human lens: Evidence of rapid insolubilizatioa of specific
polypeptides in cataract formation. Proc Natl Acad Sci USA 1978; 75:3618-3620.
31. Benedek GB: Theory of transparency of the eye. Appl Optics 1971; 10:459.
32. Spector A, Roy D: Disulfide-linked high molecular weight protein associated with
human cataract. Proc Natl Acad Sci USA 1978; 75:3244-3348.
33. Dische Z, Zil H: Studies on the oxidation of cysteine to cystine in lens proteins during
850 Taylor
cataract formation. Am J Ophthalmol 1951; 34:104-113.
34. Harding JJ: The nature and origin of the urea insoluble protein of human lens. Exp Eye
Res 1972; 13:3340.
35. Farnsworth PN, Spector A, Lozier JR, et al: The localization of a 43k polypeptide in
normal and cataractous lenses by immunofluorescence. Exp Eye Res 1981; 32:257-264.
36. Truscott R, Faull K, Augusteyn R: The identification of anthranilic acid in proteolytic
digests of cataractous human lens protein. Ophthalmic Res 1977; 9:263-268.
37. Garner WH, Garner MH, Spector A: Gamma-crystallin, a major cytoplasmic polypep-
tide disulfide linked to membrane proteins in human cataract. Biochem Biophys Res
Commun 1981; 98:439-447.
38. Kinoshita JH: Mechanisms initiating cataract formation (Proctor Lecture). Invest Oph-
thalmol Vis Sci 1974; 13:713-724.
39. Garner MH, Spector A: Selective oxidation of cysteine and methionine in normal and
senile cataractous lenses. Proc Natl Acad Sci USA 1980; 72:1274-1277.
40. Kador PF, Kinoshita JH: Diabetic and galactosaemic cataracts, in J Nugent, J Whelan
(eds): Human Cataract Formation, Ciba Foundation Symposium 106. London, Pit-
man, 1984, pp 110-131.
41. Bettelheim FA: Physical basis of lens transparency, in H Maisel (ed): The Ocular Lens:
Structure, Function, and Pathology. New York, Marcel Dekker, 1985, pp 265-300.
42. Wistow GJ, Mulders JWM, de Jong WW: The enzyme lactate dehydrogenase as a
structural protein in avian and crocodilian lenses. Nature 1987; 326:622-624.
43. Wistow G, Piatigorsky J: Recruitment of enzymes as lens structural proteins. Science
1987; 236:1554-1556.
44. Walls G, Judd H: The intraocular colour filters of vertebrates. BrJ Ophthalmol 1933;
17:641-654.
45. Duke-Elder S: General considerations: Normal age changes, in S Duke-Elder (ed):
System of Ophthalmology. St Louis, CV Mosby, 1976, Vol XI, pp 10-16.
46. Pirie A: Formation of N-formylkynurenine in proteins from lens and other sources by
exposure to sunlight. Biochem J 1971; 125:203-208.
47. van Heyningen R: The glucoside of 3-OH-kynurenine and other fluorescent com-
pounds in human lens, in The Human Lens in Relation to Cataract, Ciba Foundation
Symposium 19. Amsterdam, Associated Scientific Publishers, 1973, pp 151-168.
48. Duke-Elder S: Cataract: The pathology of cataract, in S Duke-Elder (ed): System of
Ophthalmology. St Louis, CV Mosby, 1976, Vol XI, pp 125-142.
49. Chylack LT, Ransil BJ, White 0: Classification of human senile cataractous change by
the American Cooperative Cataract Research Group (CCRG) method. III: The associa-
tion of nuclear color (sclerosis) with extent of cataract formation, age, and visual acuity.
Invest Ophthalmol Vis Sci 1984; 25:174-180.
50. Pirie A: Colour and solubility of the proteins of human cataracts. Invest Ophthalmol
1968; 7:634-650.
51. Dilley KJ: Ultraviolet light and human cataract. Nature 1975; 257:71-72.
52. van Heyningen R: Fluorescent glucoside in the human lens. Nature 1971; 230:393.
53. : Experimental studies on cataract (Proctor Lecture). Invest Ophthalmol 1976;
15:685-697.
54. Dillon J: Photolytic changes in lens proteins. Curr Eye Res 1984; 3:145-150.
55. Lerman S: Lens transparency and aging, in F Regnault, 0 Hockwin, Y Courtois (eds):
Aging of the Lens. Elsevier, Biomedical Press, 1980, pp 263-279.
56. Harding JJ, Dilley KJ: Structural proteins of the mammalian lens: A review with
emphasis on changes in development, aging, and cataract. Exp Eye Res 1976; 22:1-73.
57. Augusteyn RC: Protein modification in cataract: Possible oxidative mechanisms, in G
Duncan (ed): Mechanisms of Cataract Fornation in the Human Lens. London, Aca-
demic Press, 1981, pp 71-82.
58. Spector A: Aspects of the biochemistry of cataract, in H Maisel (ed): The Ocular Lens:
Structure, Function, and Pathology. New York, Marcel Dekker, 1985, pp 405-438.
59. Zigman S, Yulo T, Griess G: Inactivation of catalase by near-UV light. Mol CeU
 UV and the Eye 851
Biochem 1976; 11:149-154.
60. Reddy VN, Giblin FJ: Metabolism and function of glutathione in the lens, in J Nugent,
J Whelan (eds): Human Cataract Formation. Ciba Foundation Symposium 106, 1984,
pp 65-87.
61. Kalustian A, Zigman S, Sun M: Near-UV light sensitized inactivation of glutathione
reductase. Proc Am Soc Photobiol 1978, pp 112-113.
62. Bron AJ, Brown NAP: Perinuclear retrodots: A role for ascorbate in cataractogenesis.
Br J Ophthalmol 1987; 71:86-95.
63. Varma SD, Kumar S, Richards RD: Light-induced damage to ocular lens cation pump:
Prevention by vitamin C. Proc Natl Acad Sci USA 1979; 76:3504-3506.
64. Zigman S, Griess G, Yulo T, et al: Ocular protein alterations by near UV light. Exp Eye
Res 1973; 15:255-264.
65. Spector A, Garner W: Hydrogen peroxide and human cataract. Exp Eye Res 1981;
33:673-681.
66. Garner WH, Garner MH, Spector A: H202-induced uncoupling of bovine lens Na+,K+
-ATPase. Proc Natl Acad Sci USA 1983; 80:2044-2048.
67. Iwata S, Kinoshita JH: Mechanism of development of hereditary cataract in mice.
Invest Ophthalmol 1971; 10:504-512.
68. Pathak MA, Kraemer DM, Guengerich U: Formation of thymine dimers in mammalian
skin by UV radiation in vivo. Photochem Photobiol 1972; 15:177-185.
69. Zigman S, Schultz J, Yulo T: Possible roles of near UV light in the cataractous process.
Exp Eye Res 1974; 13:462-465.
70. Pitts DG, Cullen AP, Hacker PD, et al: Ocular ultraviolet effects from 295nm to 400nm
in the rabbit eye. Washington, DC, US Department of Health, Education, and
Welfare (NIOSH), 1977, pp 1-82.
71. Harding CV Jr, Susan SR, Lo W-K, et al: The strucure of the human cataractous lens,
in H Maisel (ed): The Ocular Lens: Structure, Function, and Pathology. New York,
Marcel Dekker, 1985, pp 367-404.
72. Rohrschneider W: Linsenschadigung durch ultraviolette Strahlen im Tierversuch.
Arch Ophthalmol 1936; 135:282-292.
73. Zigman S, Yulo T, Schultz J: Cataract induction in mice exposed to near UV light.
Ophthalmic Res 1974; 6:259-270.
74. Keeney AH, Rampton DV: Solar irradiation and ocular concerns. Ophthalmology
(Suppl) 1986; 93:109.
75. Wilson J: The six main causes of blindness, in International Agency for the Prevention
of Blindness, J Wilson (eds): World Blindness and Its Prevention. Oxford, Oxford
University Press, 1980, pp 1-13.
76. Pacurariu I, Marin C: Changes in the incidence of ocular disease in children and old
people. Ophthalmologica 1973; 17:289-308.
77. van Heyningen R: The human lens: I. A comparison of cataracts extracted in Oxford
(England) and Shikarpur (W Pakistan). Exp Eye Res 1972; 13:136-147.
78. Zigman S, Datiles M, Torczynski E: Sunlight and human cataracts. Invest Ophthalmol
Vis Sci 1979; 18:462-467.
79. Hiller R, Giacometti L, Yuen K: Sunlight and cataract: An epidemiologic investigation.
Am J Epidemiol 1977; 105:450-459.
80. Hiller R, Sperduto RD, Ederer F: Epidemiologic associations with cataract in the
1971-72 National Health and Nutrition Examination Survey. Am J Epidemiol 1983;
118:239-249.
81. : Epidemiologic associations with nuclear, cortical, and posterior subcapsular
cataracts. Am J Epidemiol 1986; 124:916-925.
82. Taylor HR: The environment and the lens. Br J Ophthalmol 1980; 64:303-310.
83. Hollows F, Moran D: Cataract-the ultraviolet risk factor. Lancet 1981; 2:1249-1250.
84. Mao WS, Hu TS: An epidemiologic survey of senile cataract in China. Chin Med J
1982; 95:813-818.
85. Xu J, Mao W. Zhu S, et al: A case-control study of senile cataract in lhasa, Tibet,
852 Taylor
China. Invest Ophthalmol Vis Sci (Suppl) 1987; 28:396.
86. Brilliant LB, Grasset NC, Pokhrel RP, et al: Associations among cataract prevalence,
sunlight hours, and altitude in the Himalayas. AmJ Epidemiol 1983; 118:250-264.
87. Chatterjee A: Cataract in Punjab, in The Human Lens in Relation to Cataract, Ciba
Foundation Symposium 19. Amsterdam, Associated Scientific Publishers, 1973, pp
265-279.
88. van Heyningen R: What happens to the human lens in cataract. Sci Am 1975; 233:70-
81.
89. Mann I: Culture, Race, Climate and Eye Disease. Springfield, Charles C Thomas,
1966, pp 533-534.
90. Harding JJ: Possible causes of the unfolding of proteins in cataract and a new hypothe-
sis to explain the high prevalence of cataract in some countries, in F Regnault, 0
Hockwin, Y Courtois (eds): Aging of the Lens. Elsevier, Biomedical Press, 1980.
91. HardingJJ, Rixon KC: Is diarrhoea a major cause of cataract in some tropical countries?
Metab Pediatr Ophthalmol 1981; 5:161-166.
92. Harding JJ: Cataract: Sanitation or sunglasses? Lancet 1982; 1:39.
93. Cameron LL, Auer CL, McCormick PA, et al: Association of sunlight with senile
macular and lens changes. Invest Ophthalmol Vis Sci (Suppl) 1983; 24:202.
94. Cameron LL: Association of Senile Lens and Dermal Changes with Cumulative Ultra-
violet Exposure. PhD Dissertation. Baltimore, Johns Hopkins University, 1985.
95. Cameron LL, Emmett EA, Abbey H, et al: Association of senile lens opacities with UV
and other factors. Invest Ophthalmol Vis Sci (Suppl) 1986; 27:45.
96. Tso MOM, LaPiana FG: The human fovea after sungazing. Trans AmAcad Ophthalmol
Otolaryngol 1975; 79:788-795.
97. Calkins JL, Hochheimer BF: Retinal light exposure from operation microscopes. Arch
Ophthalmol 1979; 97:2363-2367.
98. Hochheimer BF, DAnna SA, Calkins JL: Retinal damage from light. AmJ Ophthalmol
1979; 88:1029-1044.
99. Borges JM, Li ZY, Tso MOM: The cumulative effects of repeated photic exposures on
the monkey macula. Invest Ophthalmol Vis Sci (Suppl) 1987; 28:47.
100. Noell WK, Walker VS, Kang BS, et al: Retinal damage by light in rats. Invest
Ophthalmol 1966; 5:450.
101. Kuwabara T, Gom RA: Retinal damage by visible light: An electron microscopic study.
Am J Ophthalmol 1968; 79:69-78.
102. Lawwill T, Crockett S, Currier G: Retinal damage secondary to chronic light exposure.
Doc Ophthalmol 1977; 44:379402.
103. Mainster MA: Light and macular degeneration: A biophysical and clinical perspective.
Eye 1987; 1:304-310.
104. Heriot WJ: Light and the retinal pigment epithelium: The link in senile macular
degeneration, in J Cronly-Dillon, ES Rosen, J Marshall (eds): Hazards of Light,
Oxford, Pergamon Press, 1985, Vol 1, pp 187-196.
105. Weiter JJ, Delori FC, Wing GL, et al: Relationship of senile macular deeneration to
ocular pigmentation. Am J Ophthalmol 1985; 99:158-187.
106. Blumenkranz MS, Russell SR, Robey MG, et al: Risk factors in age-related maculopa-
thy complicated by choroidal neovascularization. Ophthalmology 1986; 93:552-558.
107. Sperduto RD, Hiller R, Seigel D: Lens opacities and senile maculopathy. Arch
Ophthalmol 1981; 99:1004-1008.
108. Pizzarello LD: The dimensions of the problem of eye disease among the elderly.
Ophthalmology 1987; 94:1191-1195.
109. Kupfer C: Six main causes of blindness, in International Agency for the Prevention of
Blindness, Sir J Wilson (eds): World Blindness and Its Prevention. Oxford, Oxford
University Press, 1984, Vol 2, pp 4-14.
110. US Department of Health and Human Services: Vision Research-A National Plan:
1983-1987, Vol 2/Pt 2. Bethesda, National Institutes of Health, 1983 (NIH Publication
No 83-2471).
 UV and the Eye 853
111. Maddox J: The great ozone controversy. Nature 1987; 329:101.
112. Garelik G: A breath of fresh air. Time 1987; 28:35.
113. Bowie L: The politics of ozone: Balancing interests. The Baltimore Sun Paper, Septem-
ber 27, 1987.
114. Farmer CB, Toon GC, Schaper PW, et al: Stratospheric trace gases in the spring 1986
Antarctic atmosphere. Nature 1987; 329:126-130.
115. Taylor HR, West SK: A simple system for the clinical grading of lens opacities. Lens
Res 1988; 5:175-181.
116. Sommer A: Diabetes and senile cataract. Am J Ophthalmol 1981; 91:134-135.
117. West SK, Rosenthal F, Newland HS, et al: Use of photographic techniques to grade
nuclear cataracts. Invest Ophthalmol Vis Sci 1988; 29:73-77.
118. Taylor HR, West SK: The clinical grading of lens opacities. Aust J Ophthalmol 1989;
17:81-86.
119. Bressler NM, Bressler SB, West S, et al: Senile eye changes: The grading and
prevalence of macular degeneration. Arch Ophthalmol 1989; 107:847-852.
120. Cameron LL, Vitasa BC, Lewis PG, et al: Visual assessment of facial elastosis using
photographs as a measure of cumulative ultraviolet exposure. Photodernatol 1988; pp
277-282.
121. Rosenthal FS, Bakalian AE, Phoon C, et al: Senile eye changes: Determination of
ocular exposure to ultraviolet light. Invest Ophthalmol Vis Sci (Suppl) 1987; 28:397.
122. Rosenthal FS, Phoon C, Bakalian AE, et al: The ocular dose of ultraviolet radiation in
outdoor workers. Invest Ophthalmol Vis Sci 1988; 29:649-656.
123. Cotton J: National Oceanographic and Atmospheric Agency. Personal communication.
124. Scotto J, Fears TR, Gori G: Measurements of Ultraviolet Radiation in the United States
and Comparisons with Skin Cancer Data. Bethesda, US Department of Health,
Education, and Welfare, National Cancer Institute, 1977.
125. Riegelman RK: Studying a Study and Testing a Test: How to Read the Medical
Literature. Boston, Little, Brown & Co, 1981, pp 223-242.
126. Smith AH, Waxweiler RJ, Tyroler HA: Epidemiologic investigation of occupational
carcinogenesis using a serially additive expected dose model. Am J Epidemiol 1980;
112:787-797.
127. West SK, Taylor HR: Lack of evidence for aspirin use and prevention of cataracts. Arch
Ophthalmol 1987; 105:1229-1231.
128. Peffer RS: Waterrnen. Baltimore, Johns Hopkins University Press, 1979.
129. Tucker MA, Shields JA, Hartge P, et al: Sunlight exposure as risk factor for intraocular
malignant melanoma. N Engl J Med 1985; 313:789-792.